Competitive Adsorption between Atrazine and Humic

acid onto NYEX® 1000 and its Electrochemical

Regeneration

by

Kayvan Jambakhsh

A thesis submitted under the supervision of Dr. Roberto M. Narbaitz

Submitted in Partial Fulfillment of the Requirements for the Degree of

Master of Applied Science

in

Environmental Engineering*

Department of Civil Engineering, University of Ottawa

Ottawa, Ontario, Canada.

* The Masters in Applied Sciences in Environmental Engineering is a Joint Program with

Carleton University, administrated by the Ottawa-Carleton Institute for Environmental

Engineering

© Kayvan Jambakhsh, Ottawa, Canada, 2020

ii

Abstract

Activated carbon (AC) has been extensively used as an adsorbent for the removal of target

contaminants from water. However, it has a number of limitations. First, through

competition with background organic compounds the ACs capacity to remove the target

compounds can be decreased by up 70%. This is attributed to competitive adsorption and

pore blockage by the background organics. Second, AC have a finite adsorption capacity

so granular activated carbon (GAC) in adsorbers needs to be periodically regenerated or

replaced. Third, the conventional thermal GAC regeneration technique results in the loss

of 10 to 12% of the carbon mass in each regeneration cycle. As a result, there has been

significant lab-scale research on alternative GAC methods that avoid such a significant

carbon mass loss. Unfortunately, they have not been as effective as thermal regeneration.

NYEX® is a newly-developed graphitic intercalated compound adsorbent capable of

removing different types of contaminants and which can be completely regenerated

electrochemically without a loss of adsorbent mass. The objectives of this thesis are to

determine, if like with activated carbon, the adsorption of trace target contaminants on

NYEX® is impacted by the presence of background organic compounds, and to quantify

the impact of competitive adsorption has on the electrochemical regeneration.

Experiments were conducted using atrazine, as a model toxic target contaminant, and

humic acid as a model background organic matter. Bottle-point batch isotherm tests of

single solute solutions were performed using two different approaches. The constant

adsorbate solution-varying adsorbent dose isotherms for humic acid yielded unreasonable

data for the high equilibrium liquid phase concentrations data points, it is speculated this

was caused by competition with compounds leaching from the adsorbent. The varying

iii

adsorbate solution concentration-constant adsorbent dose humic acid isotherms yielded

reasonable results and was adopted for all the remaining experiments. The humic acid

isotherm yielded Freundlich model coefficient values that were statistically the same as

those of an earlier study. A NYEX® bi-solute (atrazine and humic) isotherm experiment

showed that the humic acid decreased the atrazine adsorption capacity by an average of

37%, which is of the same order of magnitude one would expect for the same experiment

using activated carbon. Given that NYEX® lacks the large network of internal pores of AC,

pore blockage by the humic acid should not be significant yet there still was a significant

decrease in the atrazine adsorption. The literature on activated carbon adsorption isotherms

involving target contaminants and background organic matter generally does not report on

the adsorption of humic acid within these experiments, however two studies indicate that

the adsorption capacity of the background acids in only slightly impacted by the

competition. So, it was surprising that the bi-solute loading of NYEX® yielded 98% lower

humic adsorption capacities. The dominance of atrazine in the competitive adsorption tests

is attributed to its higher molar concentration in the test solutions and the higher diffusivity

of atrazine.

Batch regeneration tests were performed using a reactor with similar construction to the

reactor used by Brown et al. (2004a). Electrochemical regeneration yielded regeneration

efficiencies of up to 170% indicating the NYEX® 1000 was improved by the

electrochemical processing. Hydrogen ions could have bonded to the NYEX® and

improved the adsorption capacity of atrazine, which is a weak base, and decreased the

adsorption capacity of humic acid. Multiple regeneration cycles of the atrazine-humic acid

loaded NYEX® 1000 and atrazine-only loaded NYEX® 1000 were performed, they showed

iv

that the presence of the humic acids did not decrease the long-term atrazine adsorption

capacities. Thus, the long-term atrazine adsorption on NYEX® followed by the NYEX®

electrochemical regeneration is not impacted by the presence of humic acid, i.e., the

improvements in the NYEX® compensate for the competitive adsorption decrease in

adsorption capacity. This phenomenon should be investigated further in future research.

v

Acknowledgements

Achieving this level of academic achievement has been very challenging. Without the help

and support of several individuals, I would not have made it to this point. I want to thank

my esteemed supervisor, Dr. Roberto Narbaitz, for helping me in every step of the way

with his thoughtful support and especially for his patience and specificity during the writing

process. Also, I would like to thank Patrick D’Aoust particularly for his HPLC

troubleshooting efforts. I would also like to thank Juan Pablo Gonzalez-Galvis, who helped

train me in the use of the equipment and procedures in the laboratories.

In my everyday work, I have been fortunate to have a good group of friends, whom I am

thankful for their support, and a friendly community at the University of Ottawa. I could

not complete this degree without the support of my family. I would to thank my father,

Reza Jambakhsh, for instilling me with the confidence to pursue my Masters degree. I

would also like to thank my dear mother, Suzele Jambakhsh, for her moral support and

kind love that helped me get through the difficult times during my studies. I would also

like to thank my sister, Parisa Jambakhsh, whom provided me with humour and

thoughtfulness during certain personal struggles. Above all else, I would like to thank my

wife, Sara Mostefai, whom kept pushing me to do better and be better so that I can finish

my studies as her dutiful and loving husband.

vi

Table of Contents Abstract ................................................................................................................... ii

Acknowledgements ............................................................................................................v

List of Figures .................................................................................................................. ix

List of Tables .................................................................................................................. xi

Nomenclature ................................................................................................................. xii

Abbreviations ..................................................................................................................xv

CHAPTER 1: Introduction ..............................................................................................1

CHAPTER 2: Literature review .....................................................................................5

2.1 Adsorption mechanisms ....................................................................................... 5

2.2 Activated Carbon.................................................................................................. 6

2.3 Factors affecting adsorption ................................................................................. 7

2.3.1 Adsorbent Particle Size ................................................................................. 8

2.3.2 pH ................................................................................................................ 10

2.4 Equilibrium Adsorption Capacity Experiments ................................................. 11

2.5 Single Solute Adsorption ................................................................................... 14

2.5.1 Single solute isotherm modeling ................................................................. 15

2.5.2 Competitive Adsorption .............................................................................. 21

2.6 Activated Carbon Regeneration Methods .......................................................... 25

2.6.1 Regeneration Efficiency.............................................................................. 25

vii

2.6.2 GAC Regeneration Methods ....................................................................... 27

2.7 NYEX® and its electrochemical regeneration. ................................................... 43

2.8 Conclusions ........................................................................................................ 49

CHAPTER 3: Experimental Methods ...........................................................................50

3.1 Materials ............................................................................................................. 50

3.1.1 Adsorbent .................................................................................................... 50

3.1.2 Adsorbates................................................................................................... 51

3.2 Analytical Methods ............................................................................................ 53

3.2.1 Dissolved Organic Carbon Analysis ........................................................... 53

3.2.2 High Performance Liquid Chromatography ............................................... 54

3.3 Experimental Methods ....................................................................................... 60

3.3.1 Isotherm Experiments ................................................................................. 60

3.3.2 Regeneration Studies .................................................................................. 66

3.4 Experimental Plan .............................................................................................. 72

CHAPTER 4: Competitive Adsorption of Atrazine and its Electrochemical

Regeneration ..................................................................................................................73

4.1 Abstract .............................................................................................................. 73

4.2 Introduction ........................................................................................................ 74

4.3 Experimental Methods ....................................................................................... 77

4.3.1 Materials ..................................................................................................... 77

viii

4.3.2 Adsorption isotherm studies ....................................................................... 78

4.3.3 Electrochemical Regeneration .................................................................... 80

4.3.4 Atrazine Analysis ........................................................................................ 84

4.3.5 DOC Analysis ............................................................................................. 85

4.1 Results and Discussion ....................................................................................... 85

4.1.1 Adsorption Studies of Humic Acid ............................................................. 85

4.1.2 Adsorption Studies of Atrazine ................................................................... 90

4.1.3 Electrochemical Regeneration .................................................................... 96

4.4 Conclusions ...................................................................................................... 103

References ................................................................................................................106

CHAPTER 5: Conclusion and Recommendations .....................................................115

5.1 Conclusions ...................................................................................................... 115

5.2 Recommendations ............................................................................................ 117

References ................................................................................................................119

Appendix A ................................................................................................................134

Comments on the Atrazine Analytical Method Development .................................... 134

ix

List of Figures

Figure 2.1: Comparison between (a) an favorable (Hamdaoui and Naffrechoux 2007) and

(b) unfavorable isotherm (Ayotamuno et al. 2006) .......................................................... 15

Figure 2.2: Sample data described well by a Langmuir isotherm model (Narbaitz 2019) 17

Figure 2.3: Langmuir linearization of GAC adsorbing MTBE (Chen et al. 2010) ........... 18

Figure 2.4: Freundlich modelling of GAC adsorbing MTBE (Chen et al. 2010) ............. 21

Figure 2.5: Adsorption isotherms of atrazine in the presence of NOM on PACs ( ) at

Co,atz = 50µg/L in comparison to non-competitive isotherms ( ) at Co,ATZ = 100µg/L

(Ding 2010) ....................................................................................................................... 24

Figure 2.6: Diagram of a microwave heating system (Yagmur et al. 2017) ..................... 32

Figure 2.7: Percent phenol removal in batch tests using virgin and microwave regenerated

activated carbons (Yagmur et al. 2017) ............................................................................ 32

Figure 2.8: Cathodic regeneration versus anodic regeneration for type F-400 of GAC

loaded with phenol (Karimi-Jashni and Narbaitz 2005b) ................................................. 40

Figure 2.9: Electrochemical regeneration of phenol with GAC at 20, 67.5, and 253.3

mA/cm2 (Wang and Balasubramanian 2009) ................................................................... 41

Figure 2.10: NOM regeneration efficiency for GAC loaded at the Cincinnati WTP

(Narbaitz and McEwen 2012) ........................................................................................... 43

Figure 2.11: NYEX® 100 atrazine adsorption capacity (Ci = 6 mg/L) after numerous cycles

of anodic regeneration at 10 mA/cm2 (Brown et al. 2004a) ............................................. 45

Figure 3.1: NYEX® 1000 under SEM (Liu, et al., 2016).................................................. 50

Figure 3.2: Atrazine chemical structure (NCBI 2004) ...................................................... 53

Figure 3.3: Atrazine in H2O calibration curve measured at 230 nm ................................. 58

x

Figure 3.4: Example atrazine chromatograms: a) with the humic acid distortion; and b)

without the humic acid distortion ..................................................................................... 59

Figure 3.5: Circuit diagram (Brown et al. 2004a) ............................................................. 68

Figure 4.1: Schematic of the electrochemical reactor and circuit diagram (Brown et al.

2004a) ............................................................................................................................... 81

Figure 4.2: NYEX® 1000 adsorption of humic acid using different adsorbent doses and the

same concentration solution. ............................................................................................. 87

Figure 4.3: Alternative NYEX® 1000 adsorption of humic acid isotherms: a) using a

constant dose of NYEX® and b) using a constant Ci of humic acid ................................. 88

Figure 4.4: NYEX® 1000 single-solute atrazine adsorption isotherm .............................. 92

Figure 4.5: NYEX® adsorption isotherm for atrazine in the presence of humic acid ....... 94

Figure 4.6: Atrazine regeneration efficiency as a function of regeneration time ............. 97

Figure 4.7: Atrazine RE efficiency versus current density for NYEX® 1000 loaded with

atrazine and humic acid and NYEX® 1000 leaded with atrazine ..................................... 99

Figure 4.8: Impact of multiple regeneration cycles on the regeneration efficiency of

NYEX® loaded with atrazine and NYEX® loaded with atrazine and humic acid .......... 102

Figure 4.9: Multiple cycles of NYEX® regeneration using qee with atrazine and atrazine &

humic acid ....................................................................................................................... 103

xi

List of Tables

Table 2.1: Summary of NYEX® single synthetic organic compound adsorption and

electrochemical regeneration from the literature .............................................................. 46

Table 3.1: Atrazine sequence table in Chemstation .......................................................... 57

Table 3.2: Estimated percent error of the variables used to calculate the contaminant

loadings ............................................................................................................................. 66

Table 4.1: Humic acid isotherm model coefficients and quality of fit. ............................ 90

Table 4.2: Atrazine isotherm model coefficients and quality of fit. ................................. 91

xii

Nomenclature

C01 Initial liquid phase contaminant concentration for the first cycle (mg/L)

C02 Initial liquid phase contaminant concentration for the reloading (mg/L)

Ce Equilibrium liquid phase concentration (mg adsorbate/L)

Ce1 Equilibrium liquid phase concentration for the first cycle (mg/L)

Ce2 Equilibrium liquid phase concentration for the second cycle (mg/L)

CControl Solute concentration in the control bottle (mg adsorbate/L)

KFREUND Freundlich coefficient for the isotherm with virgin NYEX®

(mg/g/(mg/L)1/nFREUD)

KLANG Langmuir coefficient for the isotherm with virgin NYEX® (L/mg adsorbate)

M Mass of NYEX® used in the batch test (g)

M2 Mass of NYEX® used in the batch reloading test (g)

mg/L Parts of solute per million quantities of solvent (unit mass/unit volume)

n Freundlich model exponent (unitless)

pKa Acid dissociation constant

PRS Standard percent regeneration method (%)

PR2 Percent regeneration method 2 (%)

PR3 Percent regeneration method 3 (%)

xiii

µg/L Parts of solute per billion quantities of solvent (unit mass/unit volume)

q2 Solute equilibrium solid phase concentration at the end of the reloading

cycle (mg/g)

qe Equilibrium solid phase concentration (mg contaminant / g of adsorbent),

qee Regenerated adsorbent solid phase concentration that is in equilibrium with

the same liquid phase concentration as produced by the initial loading cycle

(mg/g)

qe1 Solid phase concentration of virgin adsorbent (mg/g)

qe2 Solid phase concentration of adsorbent after the readsorption cycle (mg/g)

qe11 Solid phase concentration of virgin adsorbent after the readsorption cycle

(mg/g)

qm Maximum (monolayer) adsorption of the solid phase contaminant (mg

adsorbate/g adsorbent)

qr Adsorption capacity of regenerated adsorbent after the reloading (mg/g)

V Volume of contaminated water in the equilibrium vessel (L)

V2 Volume of contaminated water used in the reloading cycle (L)

RE Regeneration efficiency (%)

𝑉𝑎𝑟(𝑞) Variance for the isotherm loading (mg adsorbate /g NYEX®)

𝜎𝐶𝐶𝑜𝑛𝑡𝑟𝑜𝑙 Standard deviation of the solute concentration of the control samples

𝜎𝐶𝑒 Standard deviation of the equilibrium liquid solute concentration

xiv

𝜎𝑉 Standard deviation of the liquid volumes

𝜎𝑀 Standard deviation of the adsorbent mass

xv

Abbreviations

AC Activated carbon

ATZ Atrazine

BET Brunauer-Emmet-Teller theory

CGAC Commercial granular activated carbon

DBP Disinfection by-products

DOC Dissolved organic carbon

DWT-PAC Powdered activated carbon made from demineralised waste tea leaves

GAC Granulated activated carbon

GIC Graphitic intercalated compound

HPLC High-performance liquid chromatography

KHP Potassium hydrogen phthalate

MTBE Methyl tert-butyl ether

N2 Nitrogen gas

NOM Natural organic matter

PAC Powdered Activated Carbon

PNP p-nitrophenol

PVC Polyvinylchloride

xvi

PZC Point zero charge

SEM Scanning electron microscope

T Temperature

TOC Total organic carbon

THM Trihalomethane

USGS United states geological survey

UV Ultraviolet light

WT-PAC Powdered activated carbon made from waste tea leaves

WTP Water treatment plant

1

CHAPTER 1: Introduction

Adsorption is a powerful contaminant separation technology that can be used to remove

many different compounds including toxic contaminants and hard to oxidize compounds.

In adsorption, the contaminants (i.e., adsorbates) are transferred from one medium (liquid

or gas) into an interface; generally, this is the surface of solid adsorbent. The most

commonly used adsorbent is activated carbon (AC), which is a highly porous manufactured

material with a very large internal surface area (~ 1000 m2/g) (Crittenden et al. 2012). The

contaminants attach themselves to sorption sites on the vast internal surface area of the

activated carbon via a number of different mechanisms. One of the key features of activated

carbon is that it is non-selective, it can be used to remove most organic compounds and a

number of heavy metals. This makes this technology very versatile. Accordingly, it is

extensively used in water treatment, industrial wastewater treatment and groundwater

remediation. A second important feature is that granular activated carbon (GAC) column

adsorbers can remove 100% of the target contaminant for a period of time (Sontheimer et

al. 1988). This is something that most competing technologies cannot achieve and it can

be extremely important in the treatment of waters containing highly toxic compounds. This

technology also has a number of limitations. First, due to the nonspecific adsorption, GAC

simultaneously removes many contaminants which compete for the available adsorption

sites (Ding 2010; Pelekani and Snoeyink 2000; Singh and Yenkie 2004). This competitive

adsorption will reduce the removal capacity of the target solutes. The extent of the

reduction will depend on the characteristics of the solutes, their size, chemical structure

and their relative concentration within the water matrix. In water treatment and

groundwater remediation applications, the naturally occurring organic matter, including

2

humic and fulvic acids, present in the water can severely decrease the adsorption capacity

of target toxic compounds (Kilduff et al. 1996; McCreary and Snoeyink 1980). It has been

theorized the natural organic matter (NOM) competes for adsorption sites and it blocks

access to some internal pores. Second, AC’s capacity to remove contaminants is limited by

the finite number of adsorption sites within the AC. As the adsorption sites on the AC

become saturated, the contaminant removal of GAC adsorbers decrease and at some point,

the GAC needs to be replaced with virgin or regenerated GAC. Due to the relatively high

cost of virgin GAC, GAC regeneration can reduce costs. However, the cost of regeneration

facilities is also quite high and it is only feasible if the carbon usage rate is high. So small

water treatment plants (WTPs) will generally purchase more virgin GAC instead of

regenerating it. Some WTPs with large GAC usage rates find it economical to have GAC

regeneration units on-site, while many WTPs have their GAC regenerated at regional

regeneration facilities, or by AC manufacturers (Chowdhury et al. 2013; Moore et al. 2003;

Narbaitz and McEwen 2012). To avoid contamination problems, water treatment plants

(WTPs) are only allowed to use virgin GAC or regenerated GAC originating from their

plant (Chowdhury et al. 2013).

Currently, thermal regeneration processes are used almost exclusively for the regeneration

of GAC, however this results in the loss of 10-12% of the GAC (AWWA and ASCE 2013;

San Miguel et al. 2001). Due to this, there has been a great deal of research on alternative

methods for the regeneration of GAC, these include chemical, biological, microwave, and

electrochemical approaches (Chan et al. 2018; Lu et al. 2011; Narbaitz and Karimi-Jashni

2009; Yagmur et al. 2017). Our group has studied cathodic electrochemical GAC

regeneration and found it to be quite effective for GAC loaded with phenol and p-

3

nitrophenol (Karimi-Jashni and Narbaitz 2005a; b; Narbaitz and Cen 1994; Narbaitz and

Karimi-Jashni 2009, 2012), but not effective in regenerating GAC originating from two

full-scale WTPs were the principal adsorbate was NOM (Narbaitz and McEwen 2012).

Dr. Roberts and his colleagues have developed an alternative adsorption-electrochemical

regeneration system using a graphitic intercalated compounds (GIC) called NYEX®

(Asghar et al. 2013; Brown et al. 2004b; a; Hussain et al. 2014; Liu et al. 2016; Mohammed

2011). Although NYEX® has few internal pores and a low surface area (~ 1-2 m2/g), it is

highly conductive (Mohammed 2011). This system has shown to be effective for removing

many different sorbed contaminants from the NYEX® and destroying them, including

NYEX® 100 loaded with atrazine (Brown et al. 2004a) and humic acid-loaded NYEX®

1000 (Asghar et al. 2013). Given that NYEX® 1000 was chosen because it was

hypothesized that its removal of trace toxic contaminants would be less impacted by the

competitive adsorption with NOM. Thus, the objectives of this thesis are: a) to investigate

the adsorption behaviour of humic acid (as a model NOM compound) and a model trace

organic contaminant on NYEX® 1000; b) to evaluate on the competitive adsorption of a

model toxic contaminant (atrazine) with a humic acids; and c) to investigate the impact of

the humic acid competitive adsorption on the NYEX®-atrazine electrochemical

regeneration efficiency. Atrazine was selected as the model trace toxic contaminant in this

study because it is one of the most commonly used herbicides in North American

agriculture and it was the most commonly detected pollutant in surface waters by the USGS

in 2001 and exceeded in some cases the concentration required (0.2 ppb) to feminize frog

populations (Duris et al. 2004; Gilliom et al. 2006). Atrazine is therefore classified as a

Group III (possibly carcinogenic to humans) by Health Canada (Health Canada 1993) and

4

banned in the European Member States in 2004 for the potential of groundwater

contamination (Health Canada 2015).

5

CHAPTER 2: Literature review

This chapter discusses the relevant literature regarding adsorption mechanisms, activated

carbon characteristics, factors affecting activated carbon adsorption, equilibrium sorption

capacities and their modeling, and methods for regenerating spent activated carbon. The

latter includes a discussion of the conventional thermal regeneration methods, research on

alternative regeneration methods (including electrochemical regeneration) as well an

alternative sorbent that can be effectively regenerated electrochemically.

2.1 Adsorption mechanisms

Adsorption is considered the mass transfer of the adsorbates (contaminants) from a solvent

(liquid or gas) to an interface where it accumulates, the interface is generally a solid

material called the adsorbent. Dissolved species in the liquid phase are transported by

diffusion to the surface of the adsorbent and for porous adsorbents further inward into the

various adsorption sites within the pores (Crittenden et al. 2012). The contaminants are

held in these sites by physical attraction to the surface or by chemical attraction. These

mechanisms are more commonly referred to as physical adsorption and chemical

adsorption. Physical adsorption is considered reversible, in most cases involving chemical

adsorption mechanisms the adsorption is not reversible since the compound and the

adsorbent surface react with one another and become irreversibly bound. Physical

adsorption is preferred to chemisorption in most cases since the adsorbents can be more

easily regenerated and possibly reused. When the contaminants are primarily chemisorbed,

the adsorbent needs to be regenerated by intensive methods, such as thermal regeneration,

or discarded. The development of many different types of adsorbents has been due to

6

attempts to expand the use of adsorbents. The most common type of adsorbent used in

water and wastewater treatment applications is activated carbon.

2.2 Activated Carbon

Activated carbon (AC) is a porous carbonaceous adsorbent that is created from

carbonaceous raw materials subjected to high temperatures in the absence of oxygen to

create a highly porous structure and to “activate” the adsorbent. The raw carbonaceous

material can be from many sources, such as various types of coal, wood, coconut shells,

hemp, etc. The material is heated to high temperatures in the absence of oxygen to

carbonise it without it being converted to ash, and then heated further while exposed to

steam, other gases or liquid chemicals to activate the pores (Van Vliet 1991). Each

manufacturer has their own proprietary manufacturing process, that differ in the

temperatures used, the temperature rise rate, and the activation agents used. Activated

carbon is comprised of turbostratic crystallites (Sontheimer et al. 1988). The crystallites

are similar to graphite, layers that are parallel to each other, except they are randomly

oriented and extensively interconnected by carbon cross links. This type of

interconnectivity leads to voids between the linkages. The voids between the crystallites

comprise the pore volume and they can vary in size (Sontheimer et al. 1988). Based on

their size pores are classified as macropores (50 nm < dp), mesopores (2 nm < dp < 50 nm),

and micropores (dp < 2 nm) (Crittenden et al. 2012). AC is a primarily a physical adsorbent

meaning it adsorbs due to physical attraction caused by nonspecific mechanisms such as

van der Waal’s forces between the positive surface of the pores on the carbon and the

negatively charged adsorbate (Crittenden et al. 2012). The van der Waal’s forces form

relatively weak bonds between both the adsorbent and the adsorbate. This attraction can be

7

reversed when exposed to the appropriate conditions such as heat. For example, in thermal

AC regeneration the spent carbon is exposed at the temperatures ranging from 800 °C and

1000 °C, the pores expand and release the adsorbate from the attached pore structure

(McLaughlin 2005). Thus, the regenerated carbon can be reused in treatment. Thermal

regeneration also regenerates AC loaded with chemisorbed contaminants, they may be

volatilized (making the adsorption site available) or carbonized on the internal surfaces of

the activated carbon (Sontheimer et al. 1988). The following subsections discusses the

factors affecting activated carbon adsorption, the determination of the equilibrium

adsorption capacities as well modeling them.

2.3 Factors affecting adsorption

The adsorption process can be affected by many factors: the raw material of the adsorbent,

size/molecular weight and charge of the adsorbate, the adsorbent pore size distribution,

overall adsorbent surface area, the surface charge, etc. (Li et al. 2002; Sontheimer et al.

1988). Factors such as the raw material of the adsorbent are important as it may impact the

activated carbon’s surface area, predominant pore size, etc. For instance, activated carbons

manufactured from liquorice residues can have a surface area of 114 m2/g while activated

carbons prepared from pistachio shells can a surface area of 1184 m2/g (Yeganeh et al.

2006). Most commercial activated carbons have surface areas in the order of 1000 m2/g.

The adsorbent surface area is particularly important as adsorption is a surface phenomenon,

thus the magnitude of the adsorption capacity is to a certain extent a function of the

adsorbent’s surface area. The following sub-sections will discuss in detail the impact of

particle size and pH on adsorption capacities, other factors such temperature can also

impact the magnitude of the adsorption.

8

2.3.1 Adsorbent Particle Size

Particle size for an adsorbent is an important factor in the rate of adsorption. Smaller

particle sizes promote an increased adsorption rate and causes the adsorbent to reach its

capacity more rapidly than if it was of a larger particle size (Chowdhury et al. 2013).

Granular activated carbon (GAC) is the larger variant of AC and can have a mean particle

size from 0.5 to 3 mm. The second classed size of AC is powdered activated carbon (PAC),

its mean particle size ranges from 20 to 50 µm. It should be noted that practically all of the

surface area of AC is within the internal pores, therefore the surface area per unit mass of

GAC and PAC is essentially the same. Accordingly, the adsorption capacity of GAC and

PAC are the same (Sontheimer et al. 1988). The key advantage of PAC is that it has much

faster kinetics, as the rate of adsorption is proportional to the inverse of the particle

diameter squared. The systems using PAC have a number of advantages and disadvantages.

PAC in water treatment is generally added at the coagulation stage and removed through

sedimentation, although sometimes the separation is via filtration (Crittenden et al. 2012).

To implement PAC additions only a PAC slurring system and a feed pump need to be

added to the treatment plant, so the capital costs of PACs addition systems are much lower

than compared to GAC contactors because no additional treatment units are required

(Chowdhury et al. 2013). This is a benefit smaller plants who have a limited budget

(Chowdhury et al. 2013). If, as in many water treatment plants, the PAC is only added

seasonally to combat algal/cyanobacterial-related taste and odour problems, the operational

cost of PAC is not high (Crittenden et al. 2012). If PAC systems are used on a continuous

basis it will lead to a larger AC consumption rate than GAC columns, which result in higher

operating costs. In addition, if high contaminant removals are required (>80% removal) the

9

PAC usage rates will be very high (Sontheimer et al. 1988). Other PAC operating concerns

include the dirtiness of the PAC slurring preparation systems, and the larger amount of

sludge wastes produced. The PAC in the sludge cannot be regenerated (Chowdhury et al.

2013).

On the other hand, GAC columns have a number of advantages. First, due to their flow

configuration the GAC achieves higher contaminant loadings that can be achieved by PAC

contact systems. Second, GAC column adsorbers can achieve up to 100% contaminant

removal for a period of time, this is critical for the removal of toxic compounds and PAC

contact systems cannot achieve 100% removal (Sontheimer et al. 1988). Third, GAC can

be economically regenerated, on-site or at regional facilities (Sontheimer et al. 1988).

However, the GAC in GAC column contactors can operate for months to years before it

becomes saturated and needs to be replaced or regenerated. The service life of the GAC

depends on the contaminant being targeted, the concentration of the contaminant, the level

of removal required, the size of the column and column system configuration (i.e., single

column operation, columns in-series, off-phase parallel columns). For the removal of

NOM, the GAC service life will be in the order of months, while for taste and odour

applications the service life is often longer than four years. For on-site regeneration

facilities to be economically feasible the GAC usage rate has to be greater than 140 000

kg/yr (Sontheimer et al. 1988). Accordingly, many plants simply replace the GAC with

virgin GAC when the capacity is reached. Due to the different contaminants, the

concentration of the contaminant, the different degrees to which they adsorb, levels of

removal required for a particular application and the economics, some water treatment

plants use PAC contact systems and others use GAC column systems. Given their greater

10

effectiveness and the possibility of regenerating the GAC, GAC column systems are

theoretically preferable.

2.3.2 pH

pH defined by the Bronsted-Lowry theory is the mixture of positively charged hydrogen

ions and negatively charged hydroxide ions in solution (Crittenden et al. 2012). This

balance of positive and negative ions greatly affects the water chemistry in a sample and

also possibly affects the mechanics of adsorption. Shi et al. (2012) studied the impact of

pH on saxitoxin adsorption on activated carbon, saxitoxin is a toxin secreted by some blue-

green algae. When the pH was below 7, the adsorption capacity of the PAC dropped

significantly. As the pH increased from 5.7 to 7.05, 8.7 and 10.7 (for a 2 hours of mixing

time), the percent saxitoxin removal increased from <10 to 48, 51, and 77%, respectively

(Shi et al. 2012). However, this pattern seems to change in the adsorption of different

compounds showing the opposite of this behaviour (Guedidi et al. 2013; Lee et al. 2007;

Ward and Getzen 1970). The inconsistency of this result has to do with AC’s point zero

charge (PZC) of the adsorbent and the adsorbate species involved. The point zero charge

is the pH where the AC is electrostatically neutral. At a pH below this level, the surface of

the AC becomes positively charged and at a level higher than this pH the surface of the AC

becomes negatively charged. The pH could also affect the adsorbate in solution, this is

because some adsorbates are transformed into ionized sister species. Ionization can occur

when the pH increases above or below the pKa value of the adsorbate and ionize to another

species. For instance, Karimi-Jashni & Narbaitz (2005b) were observed solution ionization

when adsorbing phenol. When the pH rose above the phenol pKa value of 9.99, the

adsorption capacity of the F-400 GAC dropped significantly, by pH of 13 it dropped 75%

11

(Karimi-Jashni and Narbaitz 2005b). This is because above a pH of 9.99 phenol

increasingly ionizes to phenate and has negative charge instead of a neutral charge. The

intramolecular repulsion by the negative charges causes the phenol molecules to repulse

each other and desorb off of the surface of the GAC (Cooney 1999). This would infer the

importance of controlling and/or monitoring the pH during the adsorption tests. Next, the

experiments used to quantify the maximum adsorption capacity will be discussed.

2.4 Equilibrium Adsorption Capacity Experiments

Adsorption experiments can be conducted to determine the overall capacity and determine

the favourability of adsorption of the adsorbent for a specific contaminant. This also helps

estimate the activated carbon usage rate in full-scale applications. Adsorbents only have a

limited surface area per unit mass, a particular pore size distribution and charge, therefore,

single solute equilibrium studies must be conducted to determine its maximum adsorption

for the specific contaminant-specific activated carbon combination in question.

Equilibrium studies with AC have been performed on a variety of different contaminants

that are considered hazardous to human health (Dobbs and Cohen 1980). It should be

emphasized that the adsorption capacity for a given contaminant for different types of

activated carbons will be different. In order to assess different commercial activated carbon

products, adsorption isotherms should be conducted for each of these products. The

isotherm data can be very useful to determine the quantity of a specific AC required to treat

a water that has a large concentration of a specific contaminant. Once the volume of

water/wastewater is known the mass of AC required for a PAC contact system can be

estimated using the qe (mg adsorbate/g adsorbent) in equilibrium with the system’s liquid

phase effluent contaminant concentration (Crittenden et al. 2012; Sontheimer et al. 1988).

12

In the case of a GAC column adsorbers, the mass of GAC required is based on the qe (mg

adsorbate/g adsorbent) in equilibrium with the contaminant’s feed concentration

(Crittenden et al. 2012; Sontheimer et al. 1988).

Batch equilibrium tests are generally performed to determine the maximum adsorption

capacity or loading of a target contaminant, the capacity equals the mass of contaminant

adsorbed per gram of adsorbent. As the equilibrium tests are performed at a constant

temperature they are referred to as isotherms or isotherm tests. The simplest way to conduct

isotherm tests is by the bottle-point tests at room temperature. Bottle-point isotherms tests

are experiments conducted by contacting adsorbate solutions with the adsorbent within

multiple bottles until equilibrium is reached, then the liquid solutions are separated and

their solute concentrations are measured. This final concentration is referred to as the

equilibrium liquid phase concentration (Ce). The loading or equilibrium solid phase

concentration, qe, is calculated conducting contaminant mass balances of these bottle batch

tests and rearranging it, i.e.,

𝑞𝑒 =𝑉(𝐶𝑜 − 𝐶𝑒)

𝑀

(2.1)

Where qe is the equilibrium solid phase concentration (mg contaminant / g of adsorbent),

Co is the initial (or control) liquid phase concentration (mg contaminant/ L), V is the volute

of solute solution (L), and M is the mass of the adsorbent (g). Temperature is also

significant as it can change the adsorption capacity of adsorbents, this parameter must be

kept constant during all experimentation to accurately determine the maximum adsorption

capacity at the relevant temperature.

13

There are two different methods of conducting the bottle-point isotherm tests. The first

method is to use the same solution (so Co is constant) and to change the mass of the

adsorbent in each bottle. The second approach is to use the same amount of adsorbent in

all the bottles and change the initial solute concentration of the solutions added to the

various bottles. The second approach has been used in many laboratory studies using single

adsorbate (i.e., single solute or contaminant) solutions (Asghar et al. 2013; Brown et al.

2004a). However, the constant solute / changing adsorbent dose is the most practical

approach in real applications because they may involve many solutes and one cannot

readily modify the feed concentration of all solutes and pH-impacting ions.

If the adsorbing material is not suited for a variable adsorbent dosage isotherm, such in

cases were large adsorbent doses are required, then one would use a constant adsorbent

dose and vary the concentration of the solute. In addition, the range over which the solute

concentrations can be accurately measured needs to be considered. Given the adsorption

capacity of the selected adsorbent and the adsorbent mass used in the batch experiments,

will influence the selected initial solute concentrations. Of course, the equilibrium solute

concentrations will be lower than the initial ones. So, in order to quantify the initial and

equilibrium solute concentrations with the selected analytical technique the samples may

have concentrated or diluted. Dilution would appear to be the simplest approach since

concentration would require another piece of equipment such as a centrifugal vacuum or

concentrators. In addition, if one varies the solute concentration, one needs to carry at least

one control bottle for every solute concentration used and the final concentrations in the

solutions in these bottles would need to analyzed, i.e., it requires more analytical resources.

For drinking water applications, it is better to study the adsorption using a constant solute

14

concentration and a variation in the dose since this approach saves the difficult tasks of

manipulating all the components of the water matrix. Once the isotherm data is generated

it should be plotted as qe versus Ce graph, which is referred to as the isotherm graph. Then,

the data should be modeled.

2.5 Single Solute Adsorption

Single solute equilibria isotherms tend to show a specific pattern of adsorption when

plotted. Isotherms with a convex shape (Figure 2.1a), the denomination as favourable

relates to the fact that once the mass transfer zone develops within a column it remains

constant in size as it proceeds downstream in the column with time. Unfavourable

isotherms have concave shape (Figure 2.1b), and they are considered unfavourable because

within a GAC column the mass transfer zone increases in size as it proceeds down the

column with time (Sontheimer et al. 1988). The trend shown in Figure 2.1(a) serves as a

good indicator to the adsorbent’s favourability to the adsorbate. Ayotamuno et al. (2006)

showed the adsorption of petroleum hydrocarbons from contaminated groundwater was

unfavorable adsorption, while Hamdaoui & Naffrechoux (2007) found the adsorption of

phenol was favourable. Most of the priority pollutants studied by Dobbs and Cohen (1980)

and others are adsorbed favourably. In the favourable isotherm data, it can be seen a distinct

increase in qe at the higher values of Ce to determine the maximum adsorptive capacity.

The majority of the isotherm data available is for single solutes.

15

Figure 2.1: Comparison between (a) an favorable (Hamdaoui and Naffrechoux 2007)

and (b) unfavorable isotherm (Ayotamuno et al. 2006)

2.5.1 Single solute isotherm modeling

Isotherm models are equations that describe the isotherm test data, these equations can be

used for preliminary carbon usage calculations, for comparisons among different activated

carbons, and for calibrating mass transfer models. There are many different types of models

that are appropriate depending on the circumstance (Foo and Hameed 2012; Hamdaoui and

Naffrechoux 2007). The most common models used for modelling the adsorption of

individual compounds are the Langmuir model and the Freundlich model. These models

are the standard of isotherm modelling because they only have two adjustable parameters

a

b

16

so it is relatively simple to determine the coefficient values and their values are generally

not highly correlated.

2.5.1.1 Langmuir Model

The Langmuir model, developed by Irving Langmuir in 1916, was one of the first isotherm

models (Sontheimer et al. 1988). It assumes uniformity in the adsorbent, uniformity in the

energy of adsorption of the sorption sites, adsorption at one site does not interfere with the

adsorption in the adjacent sites, and the maximum loading that can be achieved is a

monolayer. The Langmuir model equation is:

𝑞𝑒 =𝑞𝑚 ∙ 𝐾𝐿𝐴𝑁𝐺 ∙ 𝐶𝑒

1 + 𝐾𝐿𝐴𝑁𝐺 ∙ 𝐶𝑒

(2.2)

Where qm is maximum (monolayer) adsorption of the solid phase contaminant (mg

adsorbate/g adsorbent); and KLANG is the Langmuir constant (L/mg adsorbate).

Accordingly, the Langmuir describes data well that reaches a plateau value at a certain

equilibrium solid phase concentration demonstrated in Figure 2.2. Thus, if experimentally

the data reaches a plateau level of capacity, it is likely that the Langmuir model will

describe the data well.

17

Figure 2.2: Sample data described well by a Langmuir isotherm model (Narbaitz 2019)

The Langmuir model can be linearized by two different means. The first one, which is more

accurate for low concentrations, is obtained by inverting Equation 2.2, that is

1

𝑞𝑒=

1

𝐶𝑒𝑞𝑚𝐾𝐿𝐴𝑁𝐺+

1

𝑞𝑚

(2.3)

If the data is described by a straight line in a 1/qe versus 1/Ce graph, it is described well by

the Langmuir model. This can be demonstrated by the methyl tert-butyl ether (MTBE) data

in Figure 2.3. The values of the Langmuir coefficients can be obtained based on the slope

and intercepts of the plot. Alternatively, linear regression can be used to obtain the values

of the Langmuir coefficients, this approach has the added advantage that it also yields the

confidence intervals of the coefficient values.

0

50

100

150

200

250

300

350

0 10 20 30 40 50 60 70 80

qe

q(m

g co

nta

min

ant

/ g

AC

)

Ceq (mg contaminant /L)

Langmuir Isotherm

Qm

data

18

Figure 2.3: Langmuir linearization of GAC adsorbing MTBE (Chen et al. 2010)

It should be noted that the least squares regression minimizes the sum of the squares of the

errors, accordingly it gives a greater weight to the data points of larger magnitude.

Therefore, for the above Langmuir model linearization (i.e., 1/Ce and 1/qe), the regression

gives greater weight to the larger 1/qe, that is the smaller qe. Accordingly, as this

linearization may not fit the higher qe data so well, so the regressed qm value will not be

determined as accurately. The limits of the model become clear when the liquid phase

concentration is low, so the values of 1/Ce and 1/qe are high. A common alternative

approach is to use an alternative linearization of the Langmuir model, this is achieved by

multiplying by Ce to obtain

19

𝐶𝑒

𝑞𝑒=

𝐶𝑒

𝑞𝑚+

1

𝐾𝐿𝐴𝑁𝐺𝑞𝑚

(2.4)

Based on this equation, the suitability of the Langmuir model can be tested by plotting the

data in a Ce/qe versus Ce graph, if data yield a straight line the Langmuir model is adequate.

As Ce increases, qe will also increase to account for the inaccuracy at higher concentration

values. As explained by Sontheimer et al (1988) if the range of the experimental data is

rather small both linearization should work equally well, however when the experimental

data covers a large concentration range the two linearizations will yield different fits. For

such cases, they recommend using nonlinear regression.

2.5.1.2 Freundlich Model

Freundlich model is an empirical two-parameter model, it was later found to describe an

adsorbent with adsorption sites whose energy of adsorption follow an exponential

distribution (Yang 1998). As such, it describes adsorption on heterogeneous adsorbents

and activated carbon is considered to have heterogenous adsorption characteristics. The

Freundlich isotherm model is

𝑞𝑒 = 𝐾𝐹𝑅𝐸𝑈𝑁𝐷𝐶𝑒1/𝑛

(2.5)

Where KFREUND is the Freundlich constant and n is the Freundlich adsorption exponent

(unitless). The units of KFREUND depend on the units of qe and Ce, for example if the units

of Ce are mg of contaminant/L and those of qe are mg contaminant/g adsorbent then the

units of KFREUND are 𝑚𝑔 𝑐𝑜𝑛𝑡𝑎𝑚𝑖𝑛𝑎𝑛𝑡

𝑔 𝑎𝑑𝑠𝑜𝑟𝑏𝑒𝑛𝑡∙ [

𝐿

𝑚𝑔 𝑐𝑜𝑛𝑡𝑎𝑚𝑖𝑛𝑎𝑛𝑡]

1𝑛⁄

. In practice, the Freundlich model

is better suited to model heterogeneous adsorbents such as activated carbon since there are

multiple layers of adsorption sites and more than one adsorbate can adsorb onto the same

20

site (Crittenden et al. 2012). This statement contradicts one of the assumptions made by

the Langmuir equation where only one adsorbate molecule can occupy one adsorption site

(Crittenden et al. 2012). The Freundlich model has been extensively used in describing

adsorption onto activated carbon because it often models the data better than the Langmuir

model because at the applicable lower liquid phase concentrations the adsorption capacities

do not plateau (Yang 1998). The model can be linearized by applying log transformations

to yield

ln(𝑞𝑒) =1

𝑛ln(𝐶𝑒) + ln (𝐾𝐹𝑅𝐸𝑈𝑁𝐷))

(2.6)

Figure 2.4 shows a MTBE adsorption data set modelled using the linearized Freundlich

equation (Chen et al. 2010). The value of 1/n equals the slope of this line and KFREUND

equals the value of qe at the liquid phase concentration (Ce) value of 1. For the MTBE data

discussed earlier, the Freundlich model describes the higher concentrations and loadings

of MTBE better plotted the Langmuir model (Chen et al. 2010).

21

Figure 2.4: Freundlich modelling of GAC adsorbing MTBE (Chen et al. 2010)

Depending on the behaviour of the adsorbent-adsorbate combination, one model may be

better suited then the other. If higher concentrations are being explored and the qe values

reach or approach a constant level, then the Langmuir would be a more appropriate model.

2.5.2 Competitive Adsorption

In industrial wastewaters multiple contaminants are involved (Dobbs and Cohen 1980),

this is also the case for contaminated groundwater at hazardous waste sites. If activated

carbon is used when there are multiple contaminant species present in the water the

contaminants will try to compete for the adsorption sites on the pores of the adsorbent. This

phenomenon is more commonly known as multicomponent or competitive adsorption.

Most waters have a multitude of solute species that will engage in this behaviour. The

extent to which the adsorption capacity of a given target pollutant is decreased by the

22

competition depends on the adsorbability of the other compounds present and their

concentration within the water in question.

The highest concentration organic compound found in drinking water plants (and some

groundwater remediation sites) is natural organic matter (NOM). NOM is present in almost

every natural water body as well as in groundwaters. NOM is the result of decomposition

of organic matter that would be found in a lake, river or land. This includes decomposing

vegetation, microbial and algal remains, and decomposing animal matter that originate

from the water body or enters it with runoff or wastewater discharges. NOM is a

heterogeneous mixture of organic acids and is normally dominated by humic and fulvic

acids (Chowdhury et al. 2013).

Water treatment plants need to account for the NOM when treating the raw water because

of a number of reasons. Practically all water treatment plants in North America use chlorine

for primary disinfection and/or post-disinfection. If NOM is not removed before the

disinfection stage, the resulting reaction with chlorine can produce harmful disinfection

by-products (DBPs) (Chowdhury et al. 2013). Trihalomethanes (THMs) are the most

prevalent family of DBPs that are produced from chlorine reacting with organic matter.

Trihalomethanes have been proven to lead to cancer if the subject is exposed for long

periods of time to concentrations above the maximum contaminant level (CDC 2014). The

main current method of avoiding the formation of DBP formation is maximizing NOM

removal through enhanced coagulation prior to chlorination. Through higher coagulant

doses, enhanced coagulation can remove up to 50% of the NOM, but the remaining NOM

will still produce DBPs. The DBP formation may be further minimized by using alternative

disinfectants. In some cases, further NOM removal is required because enhanced

23

coagulation is not sufficiently efficient for the particular water, then AC adsorption is

sometimes implemented for this purpose (USEPA 1999)

Incorporating a GAC filter layer within the multi-media filtration step is a possibility,

however a dedicated a GAC post-filter adsorber with significant depth is required in order

to achieve large removals of NOM (Chowdhury et al. 2013). Given that GAC’s adsorption

capacity is somewhat limited for this type of applications, the GAC may have to be replaced

(or regenerated) relatively frequently (0.3 to 1 years). This makes the GAC a rather

expensive option for NOM removal and may not be the most appropriate choice

(Sontheimer et al. 1988).

In drinking water treatment, GAC is most frequently used for the removal of taste and

odour compounds, and it is occasionally used to cleanup contaminated groundwater

sources. In both cases, the target contaminants are in trace levels (μg/L or ηg/L) but these

waters still contain mg/L of NOM. And while we may not be directly interested in the

NOM removal, this NOM can impact the performance of the GAC. This is because: a)

NOM competes with the target contaminant(s) (which is present in significantly smaller

concentrations) for adsorption sites; b) the NOM is adsorbed onto the pores chemically

irreversibly; and c) the NOM can block entrance to pores (Crittenden et al. 2012; Ding

2010; Kilduff et al. 1996; Narbaitz and McEwen 2012). NOM-linked competitive

adsorption phenomena reduced the adsorption capacity of 1,1,2 trichloroethane by over

50% (Narbaitz and Benedek 1994). Ding (2010) studied the competitive adsorption of

atrazine and NOM onto various types of PACs, as shown in Figure 2.5, the atrazine

adsorption capacity of W20 PAC drops by about 85% when NOM is present (Ding 2010).

24

This may result in restricting access to unused micropores due to these molecules blocking

larger pores (Narbaitz & Benedek 1994).

Figure 2.5: Adsorption isotherms of atrazine in the presence of NOM on PACs ( ) at

Co,atz = 50µg/L in comparison to non-competitive isotherms ( ) at Co,ATZ = 100µg/L

(Ding 2010)

Regardless of the type of contaminants being remove at some point the GAC column

effluent will reach a maximum acceptable contaminants level and the GAC will have to be

replaced by virgin GAC or regenerated GAC.

25

2.6 Activated Carbon Regeneration Methods

This section will discuss the calculation of the adsorbent regeneration efficiency followed

by a discussion of alternative regeneration methods.

2.6.1 Regeneration Efficiency

Regeneration efficiency (RE) is a metric of how effective a regeneration method is in

recovering the adsorbent’s original sorption capacity. The general approach is to conduct

a batch loading tests, like those in an isotherm, with the water to be treated, followed by

the activated carbon regeneration and finally a batch reloading using the same water and

procedure. There are several ways to determine the percentage regeneration efficiency

value. The general method is calculated by determining the ratio of the adsorption capacity

of the regenerated carbon (qe2) over the adsorption capacity of the virgin carbon (Narbaitz

and Cen 1997; Sun et al. 2017). Mathematically this can be described as

𝑃𝑅𝑆 =𝑞𝑒2

𝑞𝑒1∗ 100% =

(𝐶02 − 𝐶𝑒2)𝑉/𝑀

(𝐶01 − 𝐶𝑒1)𝑉/𝑀 ∗ 100 %

(2.7)

Where PRS is the standard percent regeneration efficiency; qe2 is the solid phase

concentration of adsorbent after the reloading (readsorption) cycle (mg contaminant/g

adsorbent); qe1 is the solid phase concentration of the virgin adsorbent after the initial

loading cycle (mg contaminant/g adsorbent); C01 is the initial or control liquid phase

contaminant concentration for the first cycle (mg contaminant/L or μg contaminant/L); C02

is the initial liquid phase contaminant concentration for the reloading cycle (mg

contaminant/L or μg contaminant/L); Ce1 is the equilibrium liquid phase concentration for

the first cycle (mg contaminant/L or μg contaminant /L); Ce2 is the equilibrium liquid phase

concentration for the second cycle (mg contaminant/L or μg contaminant/L); M is the mass

26

of adsorbent used in the batch test (g); and V is the volume of contaminated water in the

equilibrium vessel (L).

There can be several alternative methods of evaluating the regeneration efficiency

(Narbaitz and Cen 1997). An alternative approach is based on using the loadings of the

regenerated GAC and the virgin GAC that correspond to the same equilibrium liquid phase

concentration (Ce1). In this approach

𝑃𝑅2 =𝑞𝑒𝑒

𝑞𝑒1∗ 100% (2.8)

Where PR2 is percentage regeneration efficiency by method 2; qee is the regenerated

adsorbent’s equilibrium solid phase concentration (mg contaminant/g adsorbent) that is in

equilibrium with the equilibrium liquid phase concentration achieved during the initial

loading cycle (Ce1). However, if one conducts the batch reloading experiment using the

same conditions as the original loading experiment (including the same mass of GAC) the

resulting equilibrium liquid phase concentration (Ce2) is unlikely to be the same as that

during the initial loading step. In order to obtain qee several reloading experiments with

slightly different initial conditions are required followed by interpolation. Accordingly, this

is a trial and error approach which requires substantial time and effort to determine qee.

Narbaitz and Cen (1997) proposed a somewhat analogous approach that requires less

experimental effort, that can be used when the adsorption is described by a Freundlich

isotherm. It uses the isotherm of the virgin GAC to estimate the loading of the virgin

adsorbent (qe11) at the equilibrium liquid phase achieved during the reloading step. The

equation for this approach is as follows

27

𝑃𝑅3 =𝑞𝑒2

𝑞𝑒11∗ 100% =

(𝐶02 − 𝐶𝑒2)𝑉/𝑀

𝐾𝐹𝑅𝐸𝑈𝐷𝐶𝑒2

1𝑛

∗ 100 % (2.9)

Where PR3 is percentage regeneration efficiency by method 3.

The second and third approaches (i.e., PR2 and PR3) seem to be more fair evaluations of

the regeneration efficiencies, because they use more consistent liquid phase concentration

basis. PR3 has been used by Narbaitz and coworkers (Karimi-Jashni and Narbaitz 2005a;

b; Narbaitz and Karimi-Jashni 2009), but by few others. This is presumably because the

higher values obtained using PRS make their regeneration methods appear to be better.

2.6.2 GAC Regeneration Methods

The conventional GAC regeneration method for GAC used in water and wastewater

treatment applications is thermal regeneration (Chowdhury et al. 2013; Crittenden et al.

2012; Moore et al. 2003). Due to some limitations there has been extensive research into

alternative GAC regeneration methods. The following subsections discuss conventional

thermal GAC regeneration as well as some of its proposed alternatives.

2.6.2.1 Conventional Thermal Regeneration

Thermal regeneration is the standard method used in regenerating spent activated carbon,

it uses both conductive and convective heat transfer to regenerate GAC. The regeneration

consists of three main steps: drying, vaporization, pyrolysis, and selective oxidation

(Chowdhury et al. 2013; Sontheimer et al. 1988; Van Vliet 1991). Each step has a gradual

escalation in temperature to obtain the desired effect. Drying usually occurs by the time the

AC reaches 200OC and as the temperature rises it also helps desorb highly volatile

adsorbates. Afterwards, the vaporization step, that occurs at 200oC to 500oC, further

28

degrades the volatile adsorbates into fragmented compounds. Then, pyrolysis occurs at

500oC to 700oC with the use of steam, it is necessary to destroy the fragmented compounds

remaining on the activated carbon. This will form some carbon residues on the surface of

the activated carbon. Selective oxidation at 700oC or higher is required by steam, carbon

dioxide and/or other oxidizing substances to oxidize the carbon residue (Van Vliet 1991).

At these conditions, the carbon residue is more reactive than the AC being treated, a small

portion of the AC will be gasified along with the carbon residue (Chowdhury et al. 2013).

Temperature at this stage is a key parameter in avoiding the carbon residue to a structure

similar to that of the activated carbon. This process is called graphitization and when the

carbon residue is treated in this manner the resulting compound will be equally as resistant

to deterioration as the activated carbon being treated. Therefore, temperatures closer to 700

OC work better than higher temperatures (Van Vliet 1991). These temperatures are used as

general guidelines since each AC manufacturer has their own process when regenerating

different types of AC. When the regeneration is complete, steam or hot gases are used to

remove the carbonaceous residue on the adsorbent. The gases carry this material as a flue

gas to be treated further. Conventional thermal regeneration is an energy intensive process.

While GAC adsorbers at water treatment plants are primarily intended for the removal of

taste and odour compounds or possibly trace levels of toxic compounds, the presence of

significantly concentrations of NOM in the water causes the NOM to be primarily loaded

with NOM. Thermal regeneration provides the intense heat necessary to volatilize NOM

and carbonize the NOM that cannot be desorbed. Regeneration efficiencies of GAC loaded

with NOM can reach up to 99% with thermal regeneration (Cannon 1993; Cannon et al.

1997; Moore et al. 2003; Narbaitz and McEwen 2012). However, the regenerated carbon

29

does lose 10-12% of its mass when thermally regenerated and may be inferior in

performance as compared to virgin GAC (AWWA and ASCE 2013; Cannon 1993; Cannon

et al. 1997; Moore et al. 2003; San Miguel et al. 2001). The value of the GAC mass losses

is due to the thermal regeneration itself plus the transportation of the adsorbent from the

adsorber to the regeneration system and then back to the adsorber.

Because of the high energy requirement and significant loss of GAC mass, there has been

significant laboratory-scale research on alternative methods of GAC regeneration in an

attempt to overcome these deficiencies. The majority of this work has involved the

regeneration of GAC loaded with model single synthetic organic compounds, such as

phenol. The alternative methods include microwave regeneration, chemical regeneration,

and electrochemical regeneration. The following subsections discuss some of key findings

of this research.

2.6.2.2 Microwave Regeneration

Microwave regeneration uses microwave radiation in the form of electromagnetic energy

where this energy is then converted into heat by a medium that can conduct a charge (i.e.,

water). Microwave research during the Second World War led to numerous developments

in the 1950s. In the 1970s, Japanese companies developed a more affordable version of the

microwave oven that was first invented in 1951 by the Raytheon Company (Yuen and

Hameed 2009). This has led to research on other applications of microwaves, of particular

interest is the use of microwaves in the regeneration of AC. Strictly speaking, this is also a

thermal regeneration process, however the different nature of the heating mechanism may

result in different regeneration efficiencies.

30

The high voltage frequency produced by microwaves create a dipole moment and changes

the direction of their pole to the opposite of the applied field (Yuen and Hameed 2009).

This creates agitation in the molecules and that agitation will generate heat. Microwave

regeneration is considered one of the most uniform methods of regeneration of GAC since

the heat transfer is occurring on a molecular level and provides an even transfer of the

microwaves into the water molecules in the AC. In contrast, conventional thermal

regeneration occurs primarily by conduction, this leads to greater heating of the outer parts

of the GAC particles and less heating of the inner parts of the GAC particles (Sun et al.

2017; Yagmur et al. 2017).

Yuen and Hameed (2009) conducted bench-scale studies involving the preparation and

regeneration of GAC. They claimed the main advantages of this approach are greater

particle interior heating, higher heating rates, greater operational control of heating, no

direct contact of the heating materials, and reduced equipment size and waste (Yuen and

Hameed 2009). Theoretically, microwave heating does not damage the carbon unlike

methods with indirect heating and consumes less energy (Yuen and Hameed 2009). Yuen

& Hameed (2009) claim regenerated carbon by this method shows is unaltered and

unaffected by the regeneration.

Other researchers such as Ania et. al. (2005) who performed a study on the effects of

thermal and microwave regeneration on the pore structure of ACs. They found that there

is a large drop in surface area (50% lower in N2 adsorption) after each microwave

regeneration cycle (Ania et al. 2005). Yagmuir et al. (2017) studied the effects of

microwave regeneration of AC loaded with different phenolic compounds. They used a

commercial activated granular activated carbon (CGAC), a powdered activated carbon

31

produced from waste tea leaves then activated by microwaves (WT-PAC) and a powdered

activated carbon produced from demineralised waste tea leaves then activated by

microwaves (DWT-PAC). The reactor configuration was very similar to a conventional

microwave oven (Figure 2.6). In order to provide venting of the gas, nitrogen was

circulated and there was an infrared pyrometer to measure the temperature within the

crucible. After production, the AC was loaded in batch tests using 400 mg/L phenol or 400

mg/L of p-nitrophenol (PNP) solutions, the AC removed 90% of these compounds. The

regeneration cycle used a microwave oven at a frequency of 2.45 GHz for 30 seconds.

Their results showed that the capacity of the CGAC, which was produced through a

conventional thermal method, to reabsorb after microwave regeneration increased to a

higher level than virgin CGAC (Figure 2.7) and the higher capacity was maintained after

subsequent regeneration cycles. Accordingly, it may be that microwave heating appears to

be able to reach some internal pores that the conventional thermal method did not manage

to initially activate. The CGAC achieved higher phenol removals than the experimental

ACs activated and regenerated using microwaves (Yagmur et al. 2017) (Figure 2.7).

However, the phenol-removal capacity of the experimental activated carbons was

maintained through several cycles of microwave regeneration, thus microwave

regeneration was 100% efficient. These regeneration efficiency results seem to contradict

the large percent losses in adsorbent area reported by Ania et al (2005).

32

Figure 2.6: Diagram of a microwave heating system (Yagmur et al. 2017)

Figure 2.7: Percent phenol removal in batch tests using virgin and microwave

regenerated activated carbons (Yagmur et al. 2017)

0

10

20

30

40

50

60

70

80

90

100

Loading 0 Regen 1 Regen 2 Regen 3 Regen 4

Rem

ova

l, %

CGAC WT-AC DWT-AC

33

The differences in phenols removals of the CGAC and the microwave prepared materials

may be more related to the AC pore size distribution of the activated carbons as they were

prepared from different raw material and activated differently. Yagmur et al. (2017)

concluded that superior removal can be achieved by preparing the activated carbon by

thermal activation process and regenerating it via microwave regeneration. However, the

comparisons were not fair. First, because the raw materials of CGAC and the other two

activated carbons were different. And secondly, there was no comparison of the impact of

thermal and microwave regeneration. So, their conclusions seem premature. The

microwave regeneration method was completed in minutes, which is a much shorter time

than the days required to regenerate the AC by conventional thermal regeneration

techniques (AWWA and ASCE 2013). On the other hand, the microwave regeneration was

conducted in the bench-scales tests and when it is scaled-up to handle the large batches of

GAC will take significantly longer than the 30 second microwave regeneration step. This

is because at water treatment plants one must also consider the time required to transport

the GAC, drain the GAC, microwave process the GAC and transport the GAC back to the

adsorber.

There is also the consideration of the regeneration of GAC loaded NOM, that adsorbs

irreversibly onto the pore structure of the AC. Zhang et. al. (2018) evaluated microwave

regeneration (T = 97 oC after 90 seconds in the microwave oven) with a 210 second

regeneration time, they observed that the maximum regeneration efficiency was less than

7%. (Zhang et al. 2018). This would mean that microwave regeneration would not be able

to desorb effectively humic acids by itself. This is a crucial requirement for any

34

regeneration method for water treatment applications since humic acids compose a

majority of all organic matter adsorbed on the GAC.

2.6.2.3 Chemical Regeneration

Chemical regeneration is the use of chemical solvents or oxidants to desorb and possibly

degrade the adsorbed compounds on the surface the adsorbent, so as to free the adsorption

sites for the next adsorption cycle. Chemical regeneration has the potential benefits of

being performed in situ or on site rather than having to invest in an onsite thermal

regeneration facility. Adsorbates can potentially be recovered during regeneration if that is

desired or profitable. The main categories of regenerants are organic and inorganic

solvents. Investigating the various mechanisms of adsorption will lead to information as to

how well GAC loaded with different compounds will be regenerated based on their

properties. It has been proven that organic solvents with high solubility are better at

regeneration than inorganic solvents with high oxidation potentials (Leng and Pinto 1996).

Also, solvents with smaller molecular weights tend to be more effective at regenerating

than solvents with higher molecular weights (Lu et al. 2011). This may be because the

smaller molecules are more effective at reaching all of the smaller pores of the adsorbents

than regenerants with a larger molecular size (Leng and Pinto 1996). Also, some

regenerants are more effective at regenerating a specific type of adsorbent than others. For

instance, sodium hydroxide was shown at being very effective at desorbing phenol as found

by Himmelstein et al. (1973). Although, investigation as to whether the residual phenolic

by-products (from the regeneration) remain adsorbed in the subsequent adsorption phase

has yet to be considered. The main by-product formed after regeneration was sodium

phenate (Himmelstein et al. 1973).

35

Lu et al. (2011) investigated the regeneration of AC loaded with two different colored dyes

(i.e., peach red and mustard yellow dyes) using a variety of chemicals. The effectiveness

of the regeneration depended on the composition and concentration of the regenerants. For

instance, 60% of acetone in 40% of water was more effective at desorption than pure

acetone. This phenomenon can be explained by the regenerant themselves, acetone as a

pure solution has a lower degree of wettability than when it is mixed with water. Wettability

is a compound’s ability in the liquid phase to adhere to a solid surface when in the presence

of other liquids. Regeneration using chemicals was impacted by a variety of factors such

as solubility (Leng and Pinto 1996). The effectiveness of the regeneration particularly

impacted the solubility of the adsorbate in the regenerant solution. The higher degree of

solubility for the dyes would have a higher driving force for desorption of the dyes. Lu et

al. (2011) evaluated the effects of regenerating the desorption efficiencies of red and yellow

dyes with over 80% desorption efficiencies for yellow dyes at 100% methanol

concentration. They claimed that it can be an acceptable alternative for thermal

regeneration If the regenerant (i.e. methanol) be used to regenerate the GAC is cost

effective (Cooney et al. 1983; Lu et al. 2011). Although, the reloading of dyes into the AC

would be challenging since the chemicals would ultimately alter the structure of the

adsorbents. Many researchers try to make the case for chemical regeneration saying that

the chemicals used are fairly volatile and can be removed with a distillation column

(Cooney et al. 1983). Although chemical regeneration may be feasibly for GAC used in

industrial wastewater applications, it is not accepted for drinking water applications

because of the potential for residual regenerant to leach from the regenerated GAC into the

treated drinking water. The potential for savings that may be achieved by using inexpensive

36

chemical regenerants is outweighed by the need for additional treatment process to

safeguard the public.

2.6.2.4 Electrochemical Regeneration Methods

Electrochemical regeneration is the process of using an electrochemical cell to regenerate

an adsorbent. Electrochemical cells can either harness chemical reactions to produce

electricity or facilitate reactions with the induction of an electrical current (Bard and

Faulkner 2001; Sillanpaa and Shestakova 2017). These reactors usually have two chambers

or cells filled with electrolyte and contain an electrode, the cells are connected through

conductive wires and electrolyte bridges. The wires and electrolyte bridges are the

conductors that complete the circuit of the cell. By applying an electrical current, the cell

hydrolyses the water and H+ & OH- ions are formed at the anode and cathode, respectively

(Asghar et al. 2013; Brown et al. 2004b; a; Hussain et al. 2014, 2015). The anode is the

oxidizing electrode and it generates very strong oxidizing agents, such as hydroxy radicals

(OHo) and chlorine (when the electrolyte contains chloride). In electrochemical GAC

regeneration systems, these ions interact with adsorbed compounds and rapidly reduce or

oxidize them. Electrochemical cells create very extreme environments in terms of pH,

anodic compartments reach pHs of approximately 2 and cathodic compartments have pHs

of approximately 12. There are two main approaches to electrochemical GAC regeneration,

they are cathodic regeneration and anodic regeneration. In principle, the regeneration

occurs by either direct oxidation on the surface of the adsorbent, by enhanced desorption,

or by enhanced desorption into the liquid phase and oxidation in the liquid phase.

Most electrochemical regeneration studies have focused on the regeneration of GAC

loaded with a model solute, the most common is phenol. Phenol is a by-product of large-

37

scale industrialized manufacturing and without treatment can end up in a body of water.

Phenol is an important target contaminant since it is widely available and very soluble in

water (Australian Government Department of Environment and Energy 2019). pH impacts

phenol (C6H5OH) adsorption in a number of different ways. First, phenol dissociate into

phenate (C6H5O-) at high pH. This is described by

C6H5OH H+ + C6H5O- K =1.2 X 10-10 (2.10)

and 𝑝𝐾𝑎 = 9.9 =[𝐻+]∙[𝐶6𝐻5O−]

[𝐶6𝐻5𝑂𝐻] (2.11)

Second, is that the adsorption capacity of phenol (and for many other compounds) is at a

maximum at pHs below the pKa, but decreases significantly at extremely low pHs and as

the pH increases above the pKa (Cooney 1999).

Narbaitz and Cen (1994) were one of the first researchers to experiment with the use of

GAC regeneration electrochemically, they evaluated it by batch loading a commercial

GAC with phenol, electrochemically regenerating it, and finally reloading it in the same

manner. They designed a reactor with the batch reactor with three horizontally oriented 7.6

cm diameter platinum sheet electrodes held 2 cm apart by a perforated plexiglass column

sections placed in an electrolyte bath. The phenol batch-loaded GAC was placed as a

monolayer on the middle electrode and the testing was conducted using very low current

densities, as low as 0.2 mA/cm2. They experimented with different types of electrolytes.

The electrolytes used were salt solutions including NaCl, NaHCO3, Na2SO4, and

CH3COONa. The best results were obtained with 1% NaCl solutions as the electrolyte, the

other salts resulted in ions that presumably ionic free radical scavengers and reduce the

free radical concentration and thus decrease the phenol oxidation (Narbaitz and Cen 1994).

38

The use of 1% NaCl electrolyte yielded a cathodic regeneration efficiency of 88% and

decreased by 2% for each subsequent regeneration cycle. Their initial regeneration

efficiency was lower than that expected for thermal regeneration, however there was no

apparent GAC mass loss after regeneration. Thermal regeneration processes can cause a

10-12% GAC mass loss and the cost of replacing that GAC with virgin material can

represent 20-40 % of the cost of GAC regeneration (AWWA and ASCE 2013; San Miguel

et al. 2001). Therefore, there may be an economic advantage in regenerating the GAC

electrochemically to avoid the cost of replacing the GAC with virgin material. There is also

the consideration of the regeneration time of GAC. Cen & Narbaitz (1994) tested various

regeneration times from 1 to 10 hours. They subsequently found that the optimal

regeneration time for GAC was 5 hours. This is considered a vast improvement from

thermal regeneration that could be up to 24 hours at the same scale. Narbaitz and Cen

(1994) studied both cathodic and anodic regeneration. The superiority of cathodic

regeneration appeared to be associated with the enhanced phenol desorption at the cathode

due to the very pH conditions it generates.

In an attempt to improve performance, Karimi-Jashni & Narbaitz (2005b) used different

electrochemical reactor configurations and they found the pH in the vicinity of the

phenolics-loaded GAC to be critical. Adding electrolyte mixing reduced the pH near the

GAC (located at the cathode) and decreased the cathodic regeneration efficiency. They also

showed a decrease in regeneration efficiency when GAC was regenerated anodically in the

electrochemical cell (Figure 2.8). The regeneration efficiencies were 30% higher when

being regenerated cathodically. The overall best regeneration efficiency of the regenerated

carbon was 80% which is considered quite low as compared to thermal regeneration which

39

generally is above 95% (Karimi-Jashni and Narbaitz 2005b; Narbaitz and McEwen 2012).

It should be noted that Karimi-Jashni and Narbaitz used Equation 2.9 to calculate the

regeneration efficiency as opposed to Equation 2.7, that yields more optimistic values, used

by Narbaitz and Cen (1994) and many others. The cathodic regeneration yielded the best

RE because of the pH at the cathode, where the loaded GAC was located, was higher than

the pKa of the adsorbed phenol which facilitated desorption (Karimi-Jashni and Narbaitz

2005b; Wang and Balasubramanian 2009). To investigate if the process could potentially

be used to regenerate a column of GAC in place, Karimi-Jashni and Narbaitz (2005b)

regenerated multiple layers of GAC particles placed on the cathode of their electrochemical

reactor. The tests showed a decrease in the regeneration efficiency with increasing bed

depth, this was presumably due to fact that only the GAC layer resting on the cathode was

exposed to the highest pH and GAC particle layers above it were exposed to lower pHs

decreasing the desorption driving force.

40

Figure 2.8: Cathodic regeneration versus anodic regeneration for type F-400 of GAC

loaded with phenol (Karimi-Jashni and Narbaitz 2005b)

Wang & Balasubramanian (2009) obtained regeneration efficiencies of nearly 100 % for

phenol-loaded GAC using anodic regeneration but they still observed a decrease in RE

with each regeneration cycle as shown in Figure 2.9. This decrease was not noted in the

Karimi Jashni & Narbaitz (2005b) study because they did not evaluate subsequent

regeneration cycles beyond the first regeneration. This long-term regeneration efficiency

decrease would lead to problems with the decrease in performance and the need for

additional virgin GAC to match performance of that of thermal regenerated GAC.

Therefore, more research must be performed on electrochemical regeneration systems to

improve their long-term performance.

Current Density = 11 A/m2

41

Figure 2.9: Electrochemical regeneration of phenol with GAC at 20, 67.5, and 253.3

mA/cm2 (Wang and Balasubramanian 2009)

Other researchers such as Berenguer et al. (2010) studied the cathodic and anodic

regeneration of phenol but obtained lower regeneration efficiencies (i.e., 70 %) than the

Wang & Balasubramanian (2009) study, which yielded REs of nearly 100%. This could be

due to a less than optimal reactor configuration as the filter press reactor’s anode

configuration was enabling the re-adsorption of oxidation by-products of phenol leading to

poor desorption of phenolics. This result appears to contradict several studies showing that

cathodic regeneration (Figure 2.8) yields somewhat higher RE than anodic regeneration

(Karimi-Jashni and Narbaitz 2005b; Narbaitz and Cen 1994; Zhang et al. 2002). Therefore,

there might be an appropriate reactor configuration and/or operating conditions that could

yield high REs for anodic regeneration. Cathodic regeneration studies were also conducted

by García-Otón et al. (2005), they used GAC loaded with toluene. At a current density of

1300 mA/cm2, versus 1.1 mA/cm2 used by Karimi-Jashni & Narbaitz (2005b), and a

regeneration time of 3 hrs, the cathodic regeneration method yielded REs above 99%

(García-Otón et al. 2005). This may be due to the change in the surface charge of the GAC

42

at higher pH levels from positive to negative making the surface repulsive to other

negatively sorbed adsorbates.

As previously discussed, GAC adsorbers at water treatment plants may remove trace

contaminants but they are primarily loaded with NOM that is present in all natural waters.

Therefore, regeneration research needs to investigate the regeneration of GAC loaded at

actual treatment plants or is loaded with NOM. There only has been limited research in this

area. Narbaitz & Karimi-Jashni (2009) showed that cathodic regeneration could achieve

regeneration efficiencies of 70 to 80% (based on the iodine numbers) for GAC from a pilot

plant column at the Britannia WTP. They speculated that these relatively high efficiencies

were due to use of the iodine number as the GAC capacity. Narbaitz & McEwen (2012)

evaluated bench-scale regeneration of NOM-loaded GAC in the lab and loaded GAC

obtained from two full-scale WTPs. They quantified the regeneration efficiency based the

TOC loadings achieved after batch reloading of the regenerated samples. They had three

interesting findings. First, they found that using the iodine number yield much larger RE

than when based on NOM (measured as TOC) reloading, thus the use of the iodine number

is not recommended. Second, the electrochemically regeneration efficiencies for field-

loaded GAC (in continuous flow GAC column adsorbers) were significantly lower (RE <

20%) than the lab batch-loaded GAC, this is presumably due to the GACs continuous

exposure to additional NOM. Thus, batch loading with NOM does not appear to be as

realistic but it is certainly easier to perform experimentally. Third, the thermal regeneration

efficiencies at these WTPs were significantly higher (RE = 85 – 100%) than those achieved

electrochemically (RE < 20%) (Figure 2.10). They speculated that cathodic regeneration

was not very effective for NOM-loaded GAC because NOM adsorbs irreversibly which

43

limits the extent of high pH driven desorption and regeneration. Thus, electrochemical

regeneration of GAC that has been loaded at water treatment plants does not seem practical.

However, for application were NOM is not the main adsorbate it may be feasible. Recently,

an effective sorption-electrochemical regeneration system has been developed using a

promising alternative sorbent, NYEX®, this research will be discussed in the next section.

Figure 2.10: NOM regeneration efficiency for GAC loaded at the Cincinnati WTP

(Narbaitz and McEwen 2012)

2.7 NYEX® and its electrochemical regeneration.

NYEX® is a graphitic intercalation compound (GIC) developed by ARVIA technologies

and Dr. Roberts’ group at the University of Manchester (Brown et al. 2004a). ARVIA

developed the early prototype of this material called NYEX® 100. It was fabricated with a

sulfuric acid treatment though the specific details of the manufacturing process are a trade

secret of the company (Asghar et al. 2013). The process enhances the surface of the

NYEX® by making it more positively charged and thus enhancing its adsorption for

44

negatively charged compounds. This can be demonstrated by comparing the adsorption of

phenol onto graphite (de Oliveira Pimenta and Kilduff 2006) with the adsorption of phenol

onto NYEX® 1000 (Asghar et al. 2015), the chemically altered graphite (NYEX® 1000)

adsorbed 1.07 times more phenol with a than unaltered graphite assuming that the type of

adsorption is irreversible (Asghar et al. 2014; de Oliveira Pimenta and Kilduff 2006). This

could be due to the type of raw material of the graphite being of higher quality and the

source material for the NYEX® was made from a lesser quality material. (Asghar 2011).

NYEX® 1000 has a much smaller total surface area than GAC (~ 1 m2/g versus ~1000 m2/g

GAC) and thus it has smaller contaminant adsorption capacities than GAC, however it is

potentially a good adsorbent due to its highly conductive nature which makes it attractive

for electrochemical regeneration. NYEX® has a much smaller surface area than GAC

because it does not have a significant pore structure. Thus, NYEX® adsorption is a much

faster process than GAC adsorption because the adsorption takes place primarily on the

surface of the NYEX® particles and it is not impacted as much by the slow diffusion of the

adsorbates into the internal pore structure.

As shown in Figure 2.11, the first anodic electrochemical regeneration of NYEX® 100

improved its atrazine sorption capacity and this high capacity was maintained through

subsequent regeneration cycles. Thus, the anodically REs were consistently above 100%.

Similar results were observed for NYEX® 1000 (another GIC) loaded with acid violet 17

dye (Mohammed 2011), and phenol (Asghar et al. 2014; Brown and Roberts 2007) (Table

2.1). It should be noted that in the NYEX® electrochemical reactors that a 1 cm deep layer

of NYEX® is placed between the anode and the membrane, which is superior to the

monolayers of GAC regenerated by the Narbaitz’s group.

45

Figure 2.11: NYEX® 100 atrazine adsorption capacity (Ci = 6 mg/L) after numerous

cycles of anodic regeneration at 10 mA/cm2 (Brown et al. 2004a)

46

Table 2.1: Summary of NYEX® single synthetic organic compound adsorption and

electrochemical regeneration from the literature

Adsorbent Adsorbate Regeneration

Time (min.)

Current

Density

(mA/cm2)

Regeneration

Efficiency (%) Reference

NYEX® 100 Atrazine 10 20 >100 (Brown et al.

2004a)

NYEX® 1000 Acid Violet 17 60 5 98 (Liu et al.

2016)

NYEX® 1000 Phenol 45 10 >100 (Asghar et

al. 2014)

NYEX® 1000 Ethanol Thiol 20 12 138 (Asghar et

al. 2019)

NYEX® 1000 Methyl

Propane Thiol 20 12 135

(Asghar et

al. 2019)

The complete electrochemical regeneration of NYEX® loaded with several different single

contaminant solutions (Table 2.1) is quite an improvement when compared to literature

results for cathodic regeneration of phenol-loaded-GAC because of several reasons. First,

Narbaitz and Karimi-Jashni (2005) found for phenol loaded GAC it achieved a maximum

cathodic regeneration efficiency of 80% versus 100% for NYEX®. This likely shows that

the structure of the GAC is being degraded over many adsorption cycles. Second, to

achieve this Narbaitz and Karimi-Jashni (2005) needed a regeneration time of 10 hours

versus the 20 minutes required for the NYEX® regeneration (Asghar et al. 2014). Third,

for the GAC studies the regeneration efficiency gradually decreased in subsequent

47

regeneration cycles, while the 100% regeneration efficiency achieved by the NYEX®

systems was sustained. On the other hand, as GAC has a much larger surface area and

much larger contaminant adsorption capacity, the GAC regenerations involved

significantly larger masses of contaminant than the NYEX® experiments. Thus, GAC’s

lower electrochemical regeneration performance efficiencies may be at least in part due to

the larger contaminant loads.

Brown et al. (2004a) explored the potential mechanisms behind the high NYEX® anodic

regeneration efficiencies. They initially thought that the hydrogen ions produced from the

electrochemical regeneration were binding to the surface of the adsorbent creating new

functional group that were improving the readsorption. However, this was apparently not

the case since loading (and reloading) the solute at different pHs had no effect on the mass

of contaminant removed from solution. Although, the regeneration of the NYEX® in this

low pH environment caused changes to its structure creating edges on the outer surface and

generating more outer surface area. This causes an increase in adsorbable surface area of

the adsorbent and thus adsorbing more compounds on the second regenerating cycle

(Brown et al. 2004a). This phenomenon would be very useful in the long-term adsorption

of target contaminants.

48

In contrast to the research using synthetic organic compounds as solutes, Asghar et al.

(2013) investigated the regeneration of NYEX® loaded with humic acid solutions.

Although these was a commercial (i.e. Aldrich) soil-extracted humic acid and are not quite

the same as aquatic-based humic substances found in surface waters, these solutions have

been often used a model NOM (McCreary and Snoeyink 1980). Asghar et al. (2013)

revealed that the anodic regeneration of the humic acid-loaded NYEX® 1000 had a RE

efficiency of 100% when the reactor was operated a current density of 10 mA/cm2 for 30

minutes. Given that humic acid are known to be quite recalcitrant and even in advanced

oxidation processes, such as ozone oxidation, only fully mineralize 10 to 20% of the humic

acid (Betancur-Corredor et al. 2016), the 100% regeneration of NYEX® was rather

surprising. This complete regeneration is not likely related to the mineralization of the

humic acid, but rather due to the enhanced desorption of the humic acid due to the humic

acid adsorbing on the surface of the NYEX® particles and the high conductivity of the

NYEX®.

One of the hypothesized mechanisms for the decreased GAC adsorption of target

compounds in the presence of NOM is that the some of the sorbed NOM blocks access to

internal pores. Accordingly, lacking a significant pore structure NYEX® may not be as

impacted as GAC by competitive adsorption involving NOM. As far as the candidate is

aware, the adsorption of target compounds on NYEX® in the presence of humics has not

been studied. Therefore, the adsorption of target compounds on NYEX® in the presence of

humic acids should be studied to determine if the lack of a significant internal pore structure

reduces the detrimental impact of competitive adsorption with humic acids.

49

2.8 Conclusions

NYEX® has been shown to have the potential to adsorb many compounds and it can be

completely regenerated by anodic electrochemical treatment. NYEX®’s high conductivity

makes it a more suitable adsorbent for electrochemical regeneration. As NYEX® lacks a

large internal pore structure it has a much smaller total surface area than GAC, so its

adsorption capacity per gram of adsorbent is significantly lower than that of GAC.

Accordingly, sorption onto NYEX® takes place on its external surface, so this should

enhance the desorption of the solute during regeneration and facilitate direct oxidation in

the surface of the adsorbent. In addition, the lack of internal pore structure may be

beneficial to NYEX®’s capacity to remove target compounds as pore blockage by

competing NOM molecules should not be a factor. This should be evaluated.

Roberts and coworkers have shown that NYEX® loaded with many different contaminants

can be effectively regenerated in their anodic regeneration systems. However, there does

not seem to be publications from other groups corroborating these results. It seems that

confirming their results, particularly the rather surprisingly high regeneration efficiencies

of NYEX® is a worthwhile area of research. In addition, as most of the NYEX® research

has involved the treatment of single contaminant solutions, the impact of competing solutes

on the regeneration efficiency of a target solute should also be studied.

50

CHAPTER 3: Experimental Methods

This chapter discusses the materials, analytical methods, experimental procedures and the

experimental plan followed in this thesis. These are described below.

3.1 Materials

3.1.1 Adsorbent

NYEX® 1000 was supplied Arvia Technologies, Runcorn, England. The 100µm to 700µm

NYEX® 1000 particles have irregular shape, are black and readily adhere to surfaces.

Figure 3.1 presents a scanning electron microscope (SEM) image of the NYEX® 1000

particles, it shows the particles have a flat plate-like shape with smooth outer surfaces and

no apparent pores. According to its manufacturer is surface area is 1 m2/g (Asghar et al.

2013).

Figure 3.1: NYEX® 1000 under SEM (Liu, et al., 2016)

51

3.1.2 Adsorbates

3.1.2.1 Humic Acid

Humic acid powder (Sigma-Aldrich, Darmstadt, Germany) was used as a model natural

organic matter (NOM). Many researchers have used this product (Asghar et al. 2013;

Kilduff et al. 1996; Li et al. 2002; Radian and Mishael 2012). A key advantage of this

product is that it is relatively inexpensive at $38.30 per 100 grams and its concentrated

form allows one to prepare humic acid solutions of different concentrations, a critical factor

in conducting isotherms with NYEX®. This particular humic acid is extracted from soils

and these compounds are the result of many processes of biological breakdown until they

have become their most recalcitrant molecular form. The Aldrich humic acid varies in

molecular size and has a molecular weight distribution ranging from 3000 to 40 000 g/mol

(Chin and Gschwend 1991). Based upon molecular weight averaging, the average

molecular becomes 4868 g/mol and thus allowing a more practical approach in quantifying

this substance (Chin and Gschwend 1991; Reid et al. 1990; Roger et al. 2010). In addition,

this was the same model NOM compound used by Asghar et al. (2013), so it’s ideal to

conduct confirmatory NYEX® isotherms and NYEX® electrochemical regeneration tests.

3.1.2.2 Atrazine

Atrazine (C8H14ClN5) was chosen as the model toxic compound for this study primarily

because Brown et al. (2004) studied its sorption on NYEX® and the electrochemical

regeneration of atrazine-loaded NYEX®, accordingly this work will help confirm their

results. Atrazine is an herbicide of the triazine class, its structure is shown in Figure 3.2. It

is used for weed control in Canada and the US (Fay et al. 1999). Research has shown that

atrazine can create hormone imbalances in both animals and humans (IARC 1991;

52

Kniewald et al. 1995). There is a possible link between atrazine and ovarian cancer.

Research done by Donna et al. (1989) found that the risk for epithelial ovarian cancer is

two to threefold as compared to unexposed woman (Donna et al. 1989). Sixty cases

between January 1974 and June 1980 were matched with 127 controls with other non-

ovarian malignancies in a region in northern Italy were the use of atrazine is extensive.

69% of all the controls showed signs of hormonally disruptive caused malignancies that

could have been herbicide related (Crosignani et al. 1984; Donna et al. 1989). Upon

ingestion of atrazine, there is 93-100% absorption into the body of rats that have a similar

biological makeup as humans (Bakke et al. 1972; Timchalk et al. 1990). Once in the body,

it can cause serious damage by deregulating the pituitary-gonadal system of hormone

regulation. Atrazine is therefore classified as a Group III (possibly carcinogenic to humans)

by Health Canada. Atrazine’s breakdown in neutral pH waters is negligible with a half life

of two years or more. Atrazine’s potential for soil contamination is high and can persist

through a full season in temperate climates (Cohen et al. 2009). Thus, the Agricultural

Canada priority scheme ranked it highest among 83 pesticides for potential of groundwater

contamination (McCrae 1991). In 2001, The United States Geological Survey ranked

atrazine as the most commonly found pesticide across the US (Duris et al. 2004; Gilliom

et al. 2006). The Europe Union Member States have banned the chemical in 2004 from the

list of authorized plant protection products due to its potential to contaminate groundwater

for long periods of time (Health Canada 2015). The Canadian guideline for the maximum

acceptable concentration of atrazine in drinking water is 5 µg/L (Health Canada 1993).

Environmental levels of atrazine in water generally range from non-detectable to 10 µg/L

53

(Duris et al. 2004; Gilliom et al. 2006). This study used analytical standard grade atrazine

(Atrazine- PESTANAL®, Sigma-Aldrich Corporation, St. Louis, MI).

Figure 3.2: Atrazine chemical structure (NCBI 2004)

3.1.1.1 Ultra-pure water

Ultrapure water was used to prepare the humic and atrazine solutions and standards,

isotherm and loading blanks, the filter rinsing steps, etc. The ultrapure water was produced

using a Milli-Q system (Millipore Corporation, Billerica, MA).

3.2 Analytical Methods

The concentrations of humic acid were quantified in terms of dissolved organic carbon

(DOC) while the concentrations of atrazine were measured using High-Performance Liquid

Chromatography (HPLC) analysis. The analytical methods followed are described below.

3.2.1 Dissolved Organic Carbon Analysis

The dissolved organic carbon (DOC) concentrations were measured following Standard

Methods’ method 5310C (APHA/AWWA/WEF 2017) using a UV-persulfate analyzer

54

(Phoenix 8000, Tekmar-Dohrmann, Cincinnati, OH). The detection limit of this device was

below 0.1 mg/L and its upper range of analysis was 20 mg/L. Therefore, concentrated

samples had to be diluted before being analysed. The DOC analyzer was standardized using

1, 2, 4, and 8 mg/L potassium hydrogen phthalate (KHP) standards created from a 1 g/L

KHP stock solution. The calibration was performed every month or when the verification

standards (used at the beginning of each analytical run) deviated by more than 10% from

the latest calibration curve. Atrazine is an organochlorinated compound so atrazine

solutions will yield a DOC concentration. However, based on its atomic composition the

carbon mass to atrazine mass ratio is 0.445, thus the DOC analyzer is not particularly

suitable to measure µg/L concentrations of atrazine. In addition, to quantify the atrazine

concentration of humic acid-atrazine solutions a chromatographic technic is necessary to

isolate the atrazine from the humic acids.

3.2.2 High Performance Liquid Chromatography

Atrazine concentrations were measured HPLC using a modified method adopted from the

U.S. Geological Surveys (USGS) (Steinheimer et al. 1990) and Agilent Technologies (Fu

2008). The HPLC analyzer (1100 series, Agilent Technologies, Waldbronn, Germany)

used a C18 Reverse Phase column 4.6 mm x 150 mm (C18 Hypersil GOLD column,

ThermoFisher Scientific, Waltham, MA, USA), and a UV detector. The analyzer had an

autosampler and the HPLC system was operated using the analyzer’s Chemstation software

interface. The optimal wavelength to study atrazine is 230 nm with a reference wavelength

of 360 nm with an isocratic mobile phase composition of 55% methanol and 45% deionised

water. The injection volume used was 100 µL and the mobile flow rate was 1 mL/min. The

detection limit of this method is approximately 2 ppb. This a suitable detection limit since

55

the lowest concentration being evaluated was approximately 5 µg/L of atrazine. The

quantification of the atrazine concentrations in the HPLC was based on the analysis of

water-based atrazine standards. Initially, an alternative atrazine analytical method was

evaluated, it involved a pre-concentration step using solid phase extraction (SPE) with

methanol elution and an injection volume of 10 µL. The detection limit of this method was

0.4 µg/L (Steinheimer et al. 1990). However, some chromatogram stability problems were

encountered in the analysis of atrazine standards. Dr. Pham suggested to try a direct

injection method, it worked well so it was adopted. It had the additional benefit that the

pre-concentration step was not necessary, but the detection limit was a bit higher (2 µg/L)

. The sensitivity of the direct injection HPLC analysis was improved by increasing the

injection volume to 100 µL. During the analysis of pure atrazine, the method worked well.

With the introduction of humic acids in order to model the competitive adsorption

behaviour, there were disruptions in the chromatograms. In order to rectify the problem,

potential solutions were examined.

These standards were prepared using a 100 mg/L atrazine stock solution in methanol that

was stored in a freezer at -20 oC. The stock solution was allowed to warm up to room

temperature (25 oC), then the atrazine standards were prepared by adding accurate volumes

of the stock solution with a micro pipettor (Single-channel mechanical pipette, VWR

Chemicals BDH®, Radnor, PA) into different size volumetric flasks ranging from 100 mL

to 1000 mL. The HPLC calibration standards ranged from 5 µg/L to 2 mg/L. The standards

were transferred from the flasks using a luer lock syringe with a 0.45 µm cellulose acetate

syringe filter, and then inserted into an HPLC autosampler vial. The vials are sealed with

septum caps and placed in the HPLC’s autosampler to initiate the analysis.

56

The sequence table for the HPLC analysis of atrazine in water used injections of water

blanks in between the injections of samples. The use of three injections of distilled water

(blanks) in between each sample was to provide a cleaning step for the column and injector

mechanism of the HPLC and to avoid cross contamination of the samples. The use of four

atrazine injections per sample was to ensure that there were at least three acceptable

chromatograms. An example of the analysis sequence is shown in Table 3.1.

.

57

Table 3.1: Atrazine sequence table in Chemstation

Sample Number of Injections

Milli-Q water blank 3

Atrazine Sample 1 4

Milli-Q water blank 3

Atrazine Sample 2 4

Milli-Q water blank 3

The atrazine analysis yields a peak that was measured in milli absorbance units (mAu) and

the peak area was calculated by the analyzer’s software, Chemstation. The calculated peak

area was plotted versus the concentration and a calibration curve was developed. Although

there was a linear response over the entire range tested, the resulting equation did not

accurately represent the lowest concentration standards, thus the calibration curve was

divided into two ranges (Figure 3.3). A high range from 200 µg/L to 2000 µg/L, and a low

range from 5 µg/L to 200 µg/L where 200 µg/L was the pivot point for both curves. Linear

equations were calculated for both ranges using Microsoft Excel®. The instrument had an

analytical standard each sample run (every 6-10 samples) to ensure quality control and was

recalibrated after replacing the reversed phase column that was reaching the end of its

column lifetime.

58

Figure 3.3: Atrazine in H2O calibration curve measured at 230 nm

The analysis of atrazine in atrazine-humic acid solutions was problematic because the

humic acid yielded a large and rather broad peak in the chromatogram at the time when the

atrazine peak appears. An example chromatogram of an injection of atrazine and humic

acid is shown in Figure 3.4. This phenomenon was likely caused by small linear organic

fractions of the humic acid. This can be confirmed since the signal from the linear organics

are strong at 230 nm. Furthermore, the phenomenon cannot be accurately predicted. For

example, in one analytical run consisting of multiple samples (each injected three or four

times), either one injection or possibly all four injections could not be used because the

phenomenon would appear. Multiple potential solutions were explored (see Appendix A)

and the following approach was selected.

y = 1.3339x + 4.3024R² = 0.9948

y = 1.4749x + 3.3671R² = 0.9985

0

500

1000

1500

2000

2500

0 200 400 600 800 1000 1200 1400 1600

Co

nce

ntr

ato

n (

µg/

L)

Area (abs)

Low end

High end

Linear (Low end)

Linear (High end)

High End

µg/L

59

To help keep the column clear of any excess humic acid that could be saturating the

column, the blanks injected in between samples were increased from three to four, the four

were two pure methanol injections followed by two (milli-Q water) blank injections. This

did not increase the reagent consumption and provided more reliable results. The only

disadvantage in using this approach is that there may be a risk that the methanol will not

suffice in clearing up the buildup of humic acid entirely and it will manifest in the first

injection point. By having the HPLC analyse four different injections of each sample, the

first can be omitted whilst the other three can be used in the data analysis.

Figure 3.4: Example atrazine chromatograms: a) with the humic acid distortion; and

b) without the humic acid distortion

a

b

60

The above method was successful as no further distortions were observed. This is possibly

due to the amount of time required to thoroughly run the column and dislodge any organics

that may be trapped inside.

3.3 Experimental Methods

The experiments conducted in this research consisted of bottle-point isotherms and

electrochemical regeneration tests. The later consisted of batch loading tests, the actual

electrochemical regeneration followed by batch reloading tests to assess the regeneration

efficiency. These are described in the subsequent subsections.

3.3.1 Isotherm Experiments

Equilibrium studies were performed to determine the adsorption characteristics and

behavior of the adsorbent when equilibrium is reached, in essence their maximum

adsorption capacity. These experiments were performed using the bottle-point method, i.e.,

each bottle is a batch contactor that generates a data point in the equilibrium (i.e. isotherm)

graph. Initially, they were performed using the same solution and different adsorbent doses

in each bottle, which is the more common approach in applied environmental research as

one generally cannot change the concentration of the contaminated water (Sontheimer et

al. 1988). Due to the unusual results produced by using this approach using NYEX® 1000,

later isotherms (indeed the majority) used the alternative bottle point-approach using

different concentration solutions in each bottle and the same NYEX® 1000 dose. The

adsorbent in question has a relatively low equilibrium time as compared with activated

carbon (Asghar et al. 2013, 2014, 2015; Mohammed 2011; Ning 2015). The equilibrium

time of NYEX® 1000 is reached after approximately 1 hour of mixing (Asghar et al. 2013).

61

Prior to conducting the bottle point experiments, the test solutions had to be prepared. First,

a 100 mg/L atrazine stock solution was prepared by weighing 10 mg of atrazine powder in

the analytical balance, it was then dissolved in methanol (Methanol ≥99.8% HiperSolv

CHROMANORM® gradient grade for HPLC, VWR Chemicals BDH®, Radnor, PA) and

mixed in a 100 mL volumetric flask until all the powder was dissolved. The stock solution

was then stored in a freezer at -20 oC until it was needed (Rodríguez-González et al. 2013).

When the stock solution was retrieved from the freezer, it was given 1 hr to reach room

temperature. This allowed the dilution to account for the change of density of methanol at

different temperatures. Once at room temperature, dilutions were made in order to achieve

the desired initial concentrations for the isotherm experiments. These solutions were

prepared by transferring the desired volume of atrazine stock solution (using a micro-

pipettor) into 1 L volumetric flasks, filing them with Milli-Q water and mixing them.

For the bottle-point isotherms conducted utilizing a constant adsorbent dose approach the

procedure was as follows. The solutions for the test were first prepared, in the case of the

atrazine isotherm in a similar fashion as the atrazine standards. The adsorbent loadings

were conducted in amber glass 250 mL bottles with Teflon®-lined screw caps, they were

labelled and Teflon® tape was wrapped around their threads (in the neck of the bottles) to

create a seal with the cap. Then the bottles were weighed with a laboratory scale (PC4400,

Mettler-Toledo Inc., Columbus, OH). This value was used later on in order to determine

the final volume of liquid in the sample. The mass of NYEX® 1000 was measured with an

analytical scale (Model#R200D, Sartorius, Goettingen, Germany), the target adsorbent

mass was 2.5 g of NYEX® 1000. The NYEX® 1000 was then loaded into the bottle then

the rest of the bottle was filled with the solution up to the neck of the bottle, and then sealed

62

with the Teflon®-lined cap. The final weight of the loaded bottle was measured using the

laboratory scale. The NYEX® 1000 dosage was kept constant for each isotherm point while

the initial concentration of the mixture varied. For the atrazine isotherm experiment the

initial solution concentrations ranged from add the atrazine concentrations used. To attain

equilibrium the contents of the bottle(s) were mixed by rotating the bottles for 1 and a half

hours in an end-over-end tumbler, it rotated at approximately 12 revolutions per minute.

Once the equilibration time was completed, the bottle(s) were removed from the tumbler

and first inspected for any leaks. The solute in the bottles were then vacuum filtered through

a cellulose acetate membrane with a nominal pore size of 0.45 µm (Whatman®, Sigma-

Aldrich, Darmstadt, Germany). The filtration process consisted of a number of steps. First,

the filter was first wetted with Mili-Q water and placed on the filter head, the 250 mL

filtering cup was put in place. To thoroughly rinse the system and the membrane, 250 ml

of Mili-Q water was first filtered through the membrane/filtration apparatus. Then 50 mL

of the sample was filtered to waste to saturate the filter holder and membrane with the

solute so that the subsequent filtrate is truly representative of the test solution. Afterwards,

the rest of the sample was filtered. Finally, the filtrate samples were analyzed to determine

the solute concentration, i.e., the solute equilibrium liquid phase concentration (Ce) of each

bottle. In addition to the bottles with NYEX® 1000, the experiments included control

bottles and blank bottles. The control samples followed the same processing as the test

samples except that NYEX® 1000 was not added to these bottles. The control bottles were

used to account for losses of solute due to reactions, volatilization and sorption onto the

bottles. There were two blank bottles that were similarly processed, one bottle was filled

with Milli-Q water and the second with Milli-Q water and the same dose of NYEX® 1000

63

as in the other bottles. The blanks were intended to quantify if there was leaching from the

bottles and/or the filtration equipment.

Then using a mass balance from the beginning of the loading until the end one can calculate

the solute equilibrium solid phase concentration (qe) achieve in each bottle. The mass

balance equation is:

𝑞𝑒 =𝑉 ∙ (𝐶𝑐𝑜𝑛𝑡𝑟𝑜𝑙 − 𝐶𝑒)

𝑀 (3.1)

Where qe is the solute equilibrium solid phase concentration (mg solute/g NYEX®); V is

the liquid volume (L); Ccontrol is the solute concentration in the control bottle (mg solute/L);

and Ce is the solute equilibrium liquid phase concentration (mg solute/L).

For the bottle-point isotherms conducted utilizing the same solution approach the

procedure was the same except that the mass of NYEX® 1000 was varied. Further details

on humic acid isotherms and the atrazine isotherms are presented in the next subsections.

3.3.1.1 Equilibrium studies with humic acid

The humic acid isotherms were conducted following essentially the same bottle-point

isotherm procedure used for the atrazine solution that was discussed in the previous sub-

section, the differences were as follows. The initial step was to create the humic acid

solutions. Humic acid (Humic acid sodium salt, Sigma-Aldrich, Darmstadt, Germany) was

mixed with Milli-Q ultrapure water for approximately 10 minutes (Millipore Corporation,

Billerica, MA) to obtain the desired initial liquid phase concentration for the specific

isotherm point. For the humic acid isotherm the initial solution concentrations ranged from

10 mg DOC/L to 25 mg DOC/L. Asghar et al. (2013) found humic acid solution reached

equilibrium with NYEX® 1000 in approximately one hour (Asghar et al. 2013). Therefore,

64

a safety factor of 30 minutes was added to this value in order to ensure equilibrium will be

reached regardless of the conditions during the experiments. Once the contacting period

and filtration steps were completed, the filtered samples the humic acid concentrations

were quantified in terms of DOC using the organic carbon analyzer.

3.3.1.2 Equilibrium procedure the atrazine plus humic acid isotherm.

The procedure for the competitive atrazine-humic acid isotherms were performed in a

similar fashion except for the following exceptions. The competitive adsorption mixture

was prepared in 2L volumetric flasks by adding different volumes of 100 mg/L atrazine

stock solution, 24 mg of humic salts and Milli-Q water. In the competitive adsorption

experiment each bottle had a different initial concentration of atrazine, the same initial

humic acid concentration and a constant mass of NYEX® 1000. After the equilibration

period, the liquid samples for analysis were extracted from the bottles using a 6 mL syringe

(Norm-Ject® Sterile Luer-Lock Syringe, VWR International LLC., Radnor, PA) and a 0.45

µm syringe filter (0.45 µm cellulose acetate syringe filter, VWR International LLC.,

Radnor, PA). The syringe was rinsed two times with sample and then the syringe filter was

attached. The syringe was then refilled and all the liquid in the syringe was filtered from

the syringe through the syringe filter to waste to allow for the filter to become saturated

and not impact the solute concentration. More sample was then loaded into the syringe and

only 4 mL was filtered to waste, the rest was loaded into an autosampler vial (11mm clear

glass crimp top vials, ThermoFisher Scientific, Waltham, MA, USA) to be analyzed by the

HPLC analyzer.

65

3.3.1.3 Isotherm Error Analysis

In order to determine if the isotherm data points are reasonable or are possibly outlier one

needs to determine the confidence limits of the data. If repeat tests are conducted the

confidence limits can be estimated based strictly on the data. If there are no replicates then

the error can be estimated based on error associated with the variables involved in the

calculation of the parameter in question. The error bars are calculated with the variance

formula and the variance of all the variables that involved in the calculation of the isotherm

loading (q). The variance (Var) of the isotherm point’s solid phase concentration is

calculated by:

𝑉𝑎𝑟(𝑞) = (𝑉

𝑀)

2

∗ 𝜎𝐶𝐶𝑜𝑛𝑡𝑟𝑜𝑙

2 + (−𝑉

𝑀)

2

∗ 𝜎𝐶𝑒

2 + (𝐶𝑜 − 𝐶𝑒

𝑀)

2

∗ 𝜎𝑉2 +

((𝐶𝑜 − 𝐶𝑒) ∗ 𝑉

𝑀2 ) ∗ 𝜎𝑀2

(3.2)

Where 𝜎𝐶𝐶𝑜𝑛𝑡𝑟𝑜𝑙 is the standard deviation of the solute concentration of the control samples;

𝜎𝐶𝑒 is the standard deviation of the equilibrium liquid solute concentration; 𝜎𝑉 is the

standard deviation of the liquid volumes; and 𝜎𝑀 is the standard deviation of the adsorbent

mass.

Equation 3.2 is an estimate that assumes that error of all the variables involved is additive,

in reality the variance calculated from experimental replicates tend to be lower because the

error in some variables may compensate for that of other variables. The standard deviation

(σ), which equals the square root of the variance, is the experimental error that can be

expected for the measurement of the given variable. Table 3.2 shows the estimated percent

error associated with each variable. The estimated percent error of the variables in the

66

isotherm equation are multiplied by the value variable to yield the standard deviation of

that variable.

Table 3.2: Estimated percent error of the variables used to calculate the contaminant

loadings

Parameters Percent Error

Co 2%

Ce 2%

M 0.02%

V 0.02%

Once the Variance has been calculated, the estimated 95% interval can be calculated using

Equation 3.3, this formula can be applied for both the positive and negative range of

possible errors. This gives the maximum variation of a data point that can be expected if

all the parameters contributed to the error in an additive fashion.

𝑞 ± 95% 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑖𝑛𝑡𝑒𝑟𝑣𝑎𝑙 = 𝑞 ± 1.96 ∗ √𝑉𝑎𝑟(𝑞) (3.3)

3.3.2 Regeneration Studies

Regeneration studies are necessary to determine how effective the adsorbent (NYEX®

1000) can be electrochemically regenerated. The following subsections will discuss the

electrochemical reactor and the procedure followed to determine the regeneration

efficiency.

67

3.3.2.1 Electrochemical reactor.

The electrochemical reactor used in this study was based on that developed by Brown et al

(2004), it is shown in Figure 3.5. It is a small batch reactor (5 x 7 cm base) used for the

regeneration of the NYEX® 1000 that is loaded in separate batch tests, the loading

procedure was very similar to that used in the equilibrium experiments. The reactor uses a

perforated plate stainless-steel cathode and a graphitic anode with a metal oxide coating.

The reactor included a 350 ion exchange membrane (Daramic 350 membrane, Daramic,

Charlotte, NC) which is located immediately beside the cathode and separates the anodic

compartment from the cathodic compartment. The reactor uses a potentiostat (Model 410,

Electrosynthesis Company Inc., East Amherst, N.Y.) converted into a galvanostat using a

5Ω resistor where the output in volts is converted into outputs of current. The NYEX® 1000

to be regenerated is placed in the anodic compartment which has a thickness of 1 cm. The

active area of contact of the NYEX® 1000 with the anode and the ion exchange membrane

is approximately 5cm by 1.5cm. This is the area used to calculate the current density. The

circuit also includes a multi-meter to determine the amount of current being inputted into

the reactor. With the active area of the anode being known (i.e., 5cm by 1.5cm), based on

this area one calculates the current density in mA/cm2. During experimentation, the impact

of several regeneration parameters was assessed. These parameters were: 1) Regeneration

time 2); Current density; and 3) The number of regeneration cycles.

68

Figure 3.5: Circuit diagram (Brown et al. 2004a)

3.3.2.2 Regeneration Efficiency Assessment Method

The regeneration assessment procedure follows three steps: 1) batch loading of the

adsorbent; 2) electrochemical regeneration of the adsorbent; and c) reloading of the

adsorbent. These will be discussed in detail below; in addition, the equation used to

calculate the regeneration efficiency is presented.

3.3.2.3 Batch loading of the adsorbent.

Firstly, the adsorbate solution was prepared. The 100 mg atrazine/L stock (in methanol)

solution must be removed from the freezer and warmed up to room temperature. The 200

µg atrazine /L solutions were prepared by diluting 2 ml of the 100 mg/L stock solution with

Milli-Q water in a 1L volumetric flask. This allows sufficient dilution of the methanol as

for it to have a negligible impact on the adsorption of atrazine. The levels of atrazine used

in this study were higher than those normally encountered in water samples (0 – 10 µg/L),

Potentiostat

5Ω Resistor

Multimeter

+

_

69

the reasons for the higher levels were: a) because adsorption reduces the contaminant

concentrations one has to start with higher concentrations; b) the study used for

comparisons (Brown et al. 2004a) used atrazine concentrations in the 0.1 to 10 mg/L range

and adsorbed onto NYEX® 100; and c) the detection limits of the analytical method

selected. In the case of humic solutions and the competitive adsorption tests using humic

acid, the humic salts were added to Milli-Q water and create 12mg DOC/L solutions.

Secondly, the 250 mL amber glass bottle (with Teflon cap) used for the loading step was

prepared, it is weighed unfilled using the laboratory scale. Third, the NYEX® 1000 was

weighed using an analytical balance, the NYEX® 1000 dose was 12 g/L. Fourth, the

NYEX® 1000 was loaded into the amber glass bottle. Fifth, the bottle was filled with the

solute leaving a small headspace to facilitate mixing. Sixth, the bottle was caped and

reweighed. The mass of solution in the bottle is calculated based on the difference in mass

between the full and empty bottles. Seventh, the bottle was placed in an end-over-end

tumbler rotating at approximately 12 rpm, to provide the necessary mixing. Eight, after 1.5

hours mixing (i.e., the equilibrium time) the bottles were removed and inspected for any

leaks. Ninth, the bottles were opened and samples were extracted using a syringe with a

0.45 µm cellulose Luer lock filter and transferred into a separate container for the chemical

analysis (e.g., HPLC autosampler vial for the atrazine analysis). This analysis yielded the

equilibrium liquid phase concentration. Tenth, the remaining contents of the amber bottle

were vacuum filtered through 0.45 µm cellulose acetate membrane filter (Whatman®,

Sigma-Aldrich, Darmstadt, Germany) to recover the NYEX® for the regeneration step.

In order to account for solute losses, as part of the loading process, a control was also

performed, that is conducting the loading process in a bottle with the solute but without the

70

NYEX®. This solute concentration along with the equilibrium liquid phase concentration

were substituted into Equation 3.4 to calculate the equilibrium solute loading on the

NYEX®.

𝑞1 =𝑉 ∙ (𝐶𝑐𝑜𝑛𝑡𝑟𝑜𝑙 − 𝐶𝑒1)

𝑀 (3.4)

Where q1 is the solute equilibrium solid phase concentration after the first loading step in

mg solute/g NYEX®; V is the liquid volume in L; Ccontrol is the solute concentration in the

control bottle in mg adsorbate/L; and Ce1 is the solute equilibrium liquid phase

concentration in mg solute/L.

3.3.2.3.1 Regeneration

The mass of NYEX® 1000 collected in the filtration step was loaded into the anodic

compartment of the reactor. A 2% NaCl electrolyte is loaded into cathode compartment

and also a small amount is used to moisten the NYEX® in the anodic compartment. This

enables greater conductivity between the adsorbent particles (Asghar et al. 2013; Brown et

al. 2004a). The volume/dimensions of the NYEX® 1000 in the reactor were required in

order to calculate the current required so as to apply the desired current density, the

dimensions were measured with a ruler. Once the desired current is calculated, the power

supply is activated and the voltage is increased or decreased based on the reading in amps

of the multimeter. When the current density is achieved, the reactor would be operated for

the selected amount of time. The electrochemical regeneration parameters that were varied

in this research were the current density and the regeneration time.

71

3.3.2.3.2 Reloading

Once the regeneration was finished, the power supply was shutoff. Then, the NYEX® was

removed from reactor and rinsed into a beaker using a wash bottle filled with a prepared

atrazine mixture. Once all the NYEX® 1000 is in the beaker, all of it is transferred into

another clean amber glass bottle to start the reloading process following essentially the

same steps as the loading process. The bottles are sealed and then weighed on the laboratory

scale. A control sample with the mixture was prepared as well. Then, the filled loading

bottles were placed in the end-over-end tumbler and rotated for 1.5 hours (the equilibration

time). After this, the bottles were removed from the tumbler and inspected for any leaks.

A sample were then extracted from each bottle with a Luer lock syringe and syringe filter

(0.45 µm cellulose acetate syringe filter, VWR International LLC., Radnor, PA) and

analyzed. The resulting concentration is referred to as the reloading equilibrium liquid

phase concentration (Ce2).

Then using a mass balance, the solute equilibrium solid phase concentration at the end of

the reloading cycle q2 was calculated, that is

𝑞2 =𝑉 ∙ (𝐶𝑐𝑜𝑛𝑡𝑟𝑜𝑙 − 𝐶𝑒2)

𝑀 (3.5)

The above loading cycle was then used to calculate the regeneration efficiency, RE, in this

study it was defined as:

𝑅𝐸(%) = 100 ∙𝑞2

𝑞1= 100 ∙

(𝑉 ∙ (𝐶𝑐𝑜𝑛𝑡𝑟𝑜𝑙 − 𝐶𝑒2)

𝑀 )

(𝑉 ∙ (𝐶𝑐𝑜𝑛𝑡𝑟𝑜𝑙 − 𝐶𝑒1)

𝑀 ) (3.6)

72

Narbaitz and Cen (1997) discussed there are alternative definitions that may provide more

equitable assessments of the regeneration efficiency and generate somewhat lower

regeneration efficiency values than equation 3.6. In this thesis, equation 3.6 was selected

as it is the most frequently used equation in the literature and thus provides a fairer

comparison with the work of others.

The adsorbent can be collected using vacuum filtration and then dried in an oven at 105 oC

overnight to remove any moisture on the adsorbent and then measured dry using the

analytical scale to calculate the final mass of adsorbent.

3.4 Experimental Plan

The objectives of this thesis are to assess the impact of humic acids on the adsorption onto

NYEX® 1000 and to assess their impact on the electrochemical regeneration of NYEX®

1000. The first set of experiments will assess the NYEX® 1000 adsorption capacity for

humic acids and atrazine. The second set of experiments will consist of NYEX® adsorption

isotherms using bi-solute (Humic acid and atrazine) solutions. Together these two set of

experiments will help quantify the impact of humic acids on the adsorption of a target toxic

compound and determine if the humic acids have as large a detrimental impact on the

adsorption of toxic compounds as that observed on activated carbon. The third group of

experiments consists of the electrochemical regeneration tests with NYEX® 1000 loaded

with atrazine and atrazine plus humic acids. The objectives of these tests are to evaluate

the impact on NYEX® 1000 regeneration of the competitive adsorption with humic acids

as well as the impact of the electrochemical process variables, including current density

and time. Finally, multiple loading and regeneration cycles were conducted to assess the

long-term feasibility of the process.

73

CHAPTER 4: Competitive Adsorption of Atrazine and its

Electrochemical Regeneration

Kayvan Jambakhsh a, Roberto M. Narbaitz a

Department of Civil Engineering, University of Ottawa, 161 Louis-Pasteur Pvt., Ottawa, ON K1N 6N5

4.1 Abstract

NYEX® is a graphitic intercalated compound (GIC) adsorbent that has been shown to

remove a number of contaminants and which can be completely regenerated

electrochemically without a loss of adsorbent mass. The objectives of this paper are to

determine, if like with activated carbon, the adsorption of trace target contaminants is

impacted by the presence of humic acid, and to quantify the impact of competitive

adsorption on the electrochemical regeneration. Batch loading tests and batch regeneration

tests were performed. The bottle-point isotherms suggest that the constant-adsorbent dose

different adsorbate concentration approach should be followed. The competitive adsorption

experiments showed that the adsorption of atrazine on NYEX® 1000 was significantly

impacted by the presence of humic acids, however surprisingly the adsorption of the humic

acids was impacted much more by the competition. Given that NYEX® 1000 lacks a

significant internal pore structure, pore blocking was unlikely to be a factor in this

competitive adsorption. Electrochemical regeneration yielded regeneration efficiencies of

up to 170% indicating the NYEX® 1000 was improved by the process. Multiple

regeneration cycles of the atrazine-humic acid loaded NYEX® 1000 showed that the

presence of the humic acids did not reduce the long-term atrazine removals.

Keywords: NYEX® 1000, atrazine, humic acid, adsorption, electrochemical regeneration

74

4.2 Introduction

Activated carbon (AC) is the most extensively used adsorbent in water treatment, industrial

wastewater treatment and groundwater remediation for the removal of toxic organic

compounds from water. Granular AC (GAC) columns can remove up to 100% of most

organic compounds and it has a relatively low cost (Sontheimer et al. 1988). The

contaminants attach themselves to sorption sites via a number of different mechanisms, the

contaminant removal is enhanced by the vast internal pore network within the AC particles

which give them a very large internal surface area (~ 1000 m2/g) (Crittenden et al. 2012;

Sontheimer et al. 1988).

Activated carbon adsorption has a number of limitations. First, due to the nonspecific

adsorption, GAC simultaneously removes many contaminants which compete for the

available adsorption sites (Ding 2010; Pelekani and Snoeyink 2000; Singh and Yenkie

2004). This competitive adsorption can significantly reduce the removal capacity of the

target contaminants (Chowdhury et al. 2013). The extent of the reduction will depend

primarily on the characteristics of the solutes, their size, chemical structure and their

relative concentration within the water matrix (Summers et al. 2012). In water treatment

and groundwater remediation applications, the highest concentration organic compounds

in the water are the naturally occurring organic matter (NOM), including humic and fulvic

acids, and they can significantly decrease the adsorption capacity of target toxic

compounds (Kilduff et al. 1996; McCreary and Snoeyink 1980). It has been theorized the

NOM competes for adsorption sites on the AC and it also blocks access to some internal

pores. Second, AC’s capacity to remove contaminants is limited by the finite number of

75

adsorption sites within the AC. As the adsorption sites on the AC become saturated, the

contaminant removal of GAC adsorbers decrease and at some point, the GAC needs to be

replaced with virgin or regenerated GAC. Due to the cost of virgin GAC most medium to

large facilities have their GAC regenerated on-site, at regional regeneration sites, or by AC

manufacturers (Chowdhury et al. 2013; Moore et al. 2003; Narbaitz and McEwen 2012).

To avoid contamination problems water treatment plants (WTPs) are only allowed to use

virgin GAC or regenerated GAC originating from their plant (Chowdhury et al. 2013).

Currently, GAC regeneration is almost exclusively performed by thermal regeneration

processes, however this results in the loss of 10-12% of the GAC (AWWA and ASCE

2013; San Miguel et al. 2001).

Due to thermal regeneration method’s significant GAC mass loss, there has been a great

deal of research on alternative GAC regeneration methods, these include biological,

chemical, electrochemical, microwave, approaches (Chan et al. 2018; Lu et al. 2011;

Narbaitz and Karimi-Jashni 2009; Yagmur et al. 2017). Our group has studied cathodic

electrochemical GAC regeneration and found it to be quite effective for GAC loaded with

phenol and p-nitrophenol (Karimi-Jashni and Narbaitz 2005a; b; Narbaitz and Cen 1994;

Narbaitz and Karimi-Jashni 2009, 2012), and they hypothesized that the main cathodic

regeneration mechanism was high pH enhanced desorption at the cathode. However,

cathodic regeneration was not effective in regenerating GAC originating from two full-

scale WTPs at which the principal adsorbate was NOM (Narbaitz and McEwen 2012), the

lack of effectiveness was attributed to the principally irreversible chemisorption of NOM

on the GAC.

76

Dr. Roberts and his colleagues have developed an alternative adsorption-electrochemical

regeneration system using a graphite intercalation compounds (GIC) called

NYEX®(Asghar et al. 2013; Brown et al. 2004b; a; Hussain et al. 2014; Liu et al. 2016;

Mohammed 2011). Although NYEX® has few internal pores and a low surface area (~ 1-

2 m2/g), it is highly conductive (Mohammed 2011). This system has shown NYEX® to be

effective for removing many different contaminants and it can be electrochemically

regenerated. In addition, the kinetics of adsorption on NYEX® are much faster than those

for GAC, likely due to NYEX®’s lack of a significant pore structure. The contaminants

studied included atrazine (Brown et al. 2004a), phenol (Asghar et al. 2014; Brown and

Roberts 2007) and humic acid (Asghar et al. 2013). Given that NYEX® 1000 lacks the

large internal pore structure of GAC, it is hypothesized that the removal of trace toxic

contaminants by NYEX® would be less impacted than GAC by competitive adsorption

with NOM. Thus, the objectives of this thesis are: a) to investigate the adsorption behaviour

of humic acid (as a model NOM compound) and atrazine, as a model trace organic

contaminant, on NYEX® 1000; b) to evaluate on the competitive adsorption of atrazine

with a humic acids; and c) to investigate the impact of the humic acid competitive

adsorption on the NYEX®-atrazine electrochemical regeneration efficiency. Atrazine was

selected as the model trace toxic contaminant in this study because it is one of the most

commonly used herbicides in North American agriculture. Atrazine was the most

commonly detected pollutant in surface waters by the USGS in 2001 and in some cases it

exceeded the concentration required (0.2 μg/L) to feminize frog populations (Duris et al.

2004; Gilliom et al. 2006). Atrazine has been classified as a Group III (possibly

carcinogenic to humans) by Health Canada (Health Canada 1993) and it is banned in the

77

European Member States in 2004 due to its potential for groundwater contamination

(Health Canada 2015).

4.3 Experimental Methods

4.3.1 Materials

The NYEX® 1000 was supplied by Arvia Technologies in the form of a dry powder.

NYEX® 1000 has a very low surface BET surface area (~ 1 m2/g) as compared to

commercial AC which generally has surface areas in the 800-1500 m2/g range (Cannon

1993; Hamdaoui and Naffrechoux 2007; Randtke and Snoeyink 1983; Yeganeh et al.

2006). The average particle size is 484 µm and can range from 100-700 µm (Asghar et al.

2014). The material was very light and had a material density of 0.89 g/cm3 (Asghar et al.

2013). Due to this material’s high conductivity, it would accumulate any charge that is

present in its environment and would thus statically adhere onto surfaces during laboratory

experiments (Asghar et al. 2013). This phenomenon was observed to be more predominant

with the smaller particles which require less charge in order to become electrostatic-

adhesive as they weigh less. Prior to the testing pre-washing of the NYEX® was attempted,

however because of the electro-adhesive characteristics (particularly of the smaller

particles) it led to particle losses. Accordingly, it was decided to perform the experiments

using the as-received NYEX®. Similarly, recovering all the NYEX® from one experiment

to use in the next experiment was challenging, as the NYEX® particles will adhere to the

containers and equipment used. Also, since the smallest NYEX® particles are the most

adhesive they may not be recovered and as they also have the largest surface area per unit

mass, Ning (2015) speculated that this may impact the adsorption capacity NYEX®.

78

Atrazine (PESTANAL®, VWR, St. Louis, MI) used in these experiments had a purity of

99.9%. For easier dilution all the aqueous atrazine standards and tests solutions were

prepared using a methanol-based 100 mg atrazine/L stock solution. Humic acid solutions

were used to model/simulate the NOM found in natural waters. The humic acid solutions

were prepared using a soil-derived humic acid sodium salt purchased from Sigma-Aldrich

(St. Louis, MI). The dissolved organic carbon concentration to humic acid mass

concentration of the humic acid solutions was 0.67. The humic acid product has a wide

range of molecular weights, the average molecular is 4868 g/mol (Chin and Gschwend

1991; Reid et al. 1990; Roger et al. 2010).

4.3.2 Adsorption isotherm studies

Bottle-point isotherm tests were conducted to determine the isotherm characteristics for

both humic acids and atrazine as single solutes and of bi-solute mixtures of atrazine and

humic acid. Initially, the bottle-point tests were performed using the same solution and

different adsorbent doses in each bottle, which is the more common approach in applied

environmental research as one generally cannot change the concentration of the

contaminated water (Sontheimer et al. 1988). Due to the unusual results produced by using

this approach with NYEX®, later isotherms (indeed the majority) were performed by the

bottle point-approach using different concentration solutions in each bottle and the same

NYEX® dose. The tests were conducted using 250 ml glass bottles with Teflon® lined

screw caps. Each isotherm included a distilled water blank bottle and a distilled water plus

NYEX® blank bottle that were mixed along side the samples, these tests were performed

to quantify if there was leaching from the NYEX® and from the bottles. 250 mL control

bottles were filled with the adsorbate solution without any adsorbent in order to observe

79

any change in the adsorbate concentration during the loading, they were processed along

side the bottles containing NYEX®. The humic acid adsorption kinetics onto NYEX® are

very fast, Asghar et al. (2013) found that equilibrium was achieved in 50 minutes. This is

much faster than the adsorption of humic acids by GAC which could take one week or 168

hours to complete (McCreary and Snoeyink 1980). Given that humic acids, which have a

complex molecular structure and a large molecular weight, achieves adsorption

equilibrium on NYEX® within 50 minutes, then atrazine, which has a simpler molecular

structure and smaller size, should be adsorbed within 50 minutes or faster. Therefore, as a

precaution we adopted a mixing time of 1h30 minutes. These bottles were sealed and

placed in an end-over-end tumbler rotating at 12 rpm to provide the necessary mixing for

1h 30 minutes to reach equilibrium. The equilibration time was selected based on the results

of Asghar et al. (2013). Afterwards, the contents of the individual bottles were filtered

using a vacuum filtration system with a 0.45 um cellulose acetate membrane filters

(Whatmann®, Sigma-Aldrich, Darmstadt, Germany). The filtrate was then analyzed for its

atrazine and/or humic acid concentrations, the latter was quantified in terms of the

dissolved organic carbon (DOC) concentration. For the atrazine isotherms the initial

atrazine solutions used in the isotherms ranged from 20 μg/L to 1000 μg/L, for the humic

acid isotherm the initial DOC concentrations ranged from 8 mg DOC/L to 16 mg DOC/L.

The levels of atrazine used in this study were higher than those normally encountered in

water samples (0 – 10 µg/L), the reasons for the higher levels were: a) because adsorption

reduces the contaminant concentrations one has to start with higher concentrations; b) the

study used for comparisons (Brown et al. 2004a) used atrazine concentrations in the 0.1 to

80

10 mg/L range and adsorbed onto NYEX® 100; and c) the detection limits of the analytical

method selected.

Then using a mass balance from the beginning of the loading step until the end of it one

can calculate the solute equilibrium solid phase concentration (qe) achieve in each bottle.

The mass balance equation reduces to

𝑞𝑒

=𝑉 ∙ (𝐶𝑐𝑜𝑛𝑡𝑟𝑜𝑙 − 𝐶𝑒)

𝑀 (4.1)

Where qe is the solute equilibrium solid phase concentration (mg solute/g NYEX®); V is

the liquid volume (L); Ccontrol is the solute concentration in the control bottle (mg solute/L);

and Ce is the solute equilibrium liquid phase concentration (mg solute/L). The isotherm

data will be modelled with the Langmuir and Freundlich models by using the linearized

versions of these models (Sontheimer et al. 1988) and regressed using Microsoft Excel®’s

linear regression tool.

4.3.3 Electrochemical Regeneration

A small (5 x 7 cm base) polyvinylchloride (PVC) electrochemical batch reactor (Figure

4.1) was constructed based on the specifications of Brown et al. (2004a). The reactor

consisted of parallel divided cells separated by a microporous Daramic 350 membrane

(Daramic, Charlotte, NC). The perforated stainless-steel cathode within the anodic

compartment was placed immediately besides the membrane. While the graphitic anode

with a mixed metal oxide coating (supplied by Alfa Aesar, Tewksbury, MA) was located

in the far side of the anodic compartment (i.e., 10 mm from the membrane). For the

regeneration, the loaded NYEX® was place in the anodic compartment of the

electrochemical reactor and lightly wetted with a 2% w/w NaCl solution. As the NYEX®

81

was not always uniformly loaded the active area of the anode varied, typically it was 7.5

cm2. The reactor uses a potentiostat (Model 410, Electrosynthesis Company Inc., East

Amherst, N.Y.) converted into a galvanostat using a 5Ω resistor where the output in volts

is converted into outputs of current.

Figure 4.1: Schematic of the electrochemical reactor and circuit diagram (Brown et al.

2004a)

The effectiveness of electrochemical regeneration was assessed based on the following

three sets of experiments: 1) batch loading of the adsorbent; 2) electrochemical

regeneration of the adsorbent; and c) reloading of the adsorbent. These steps are required

to calculate the regeneration efficiency. The procedure to assess the regeneration efficiency

of NYEX® loaded with atrazine and that for humic acids plus atrazine solutions was the

same, only the initial solute concentrations and the analytical methods used to quantify

their concentrations differed. The batch loading was performed in the same fashion as the

bottle-point isotherms, 3 g of NYEX® was mixed with 250 mL of µg/L atrazine solution

Potentiostat

5Ω Resistor

Multimeter

+

_

82

for 1 h 30 min. After equilibrium is reached, a sample was first extracted for the atrazine

analysis using a luer lock syringe (Norm-Ject® Sterile Luer-Lock Syringe, VWR

International LLC., Radnor, PA)) with a syringe filter (0.45 µm cellulose acetate syringe

filter, VWR International LLC., Radnor, PA). Then the NYEX® was recovered through

vacuum filtration through a 47 mm diam. 0.45 µm cellulose acetate membrane filter

(Whatman®, Sigma-Aldrich, Darmstadt, Germany). Next, the electrochemical regeneration

was performed. The recovered loaded NYEX® (approximately 50% dry solids) adsorbent

was placed into the anodic compartment of the electrochemical reactor. Then, the adsorbent

was lightly wetted with a 2% NaCl solution to enhance conductivity within the reactor

(Brown et al. 2004b; a; Karimi-Jashni and Narbaitz 2005a). The cathode compartment was

loaded with the same NaCl solution until it was level with the NYEX® bed in the anodic

compartment. A DC current was applied based on the active area of the anode, it varied

from 5-40 mA/cm2 with regeneration times ranging from 5 to 50 min. The bed mass was

not mixed except by the bubbles that were the by-product of the electrolysis. After

completing the electrochemical regeneration, the regenerated NYEX® was transferred from

the anodic compartment to a 250 ml glass bottle used for the reloading. Due small size of

the anodic compartment and the electrostatic characteristics of the NYEX® this was a

problematic step, a wash bottle filled with the reloading solution was used to facilitate the

transfer of the NYEX®. To reduce these problems for future research it is recommended

that adsorption and electrochemical regeneration be performed in the same reactor, Roberts

and colleagues developed such a reactor for the Arvia® Process (Mohammed 2011).

Finally, the reloading step was performed using the same procedure as the loading step,

i.e., contacting the regenerated NYEX® with 250 mL of 200 µg/L atrazine solution for 1 h

83

30 min. At the end of the loading, the solutions were filtered to obtain samples for the

atrazine analysis and to recover the NYEX®. The NYEX® was placed in an aluminum

sampling dish and dried in 105oC oven to evaporate any excess moisture that would

interfere with the final dry weighted value.

The electrochemical regeneration efficiency was calculated using

𝑅𝐸(%) = 100 ∙𝑞2

𝑞1= 100 ∙

(𝑉 ∙ (𝐶𝑐𝑜𝑛𝑡𝑟𝑜𝑙 − 𝐶𝑒2)

𝑀)

(𝑉 ∙ (𝐶𝑐𝑜𝑛𝑡𝑟𝑜𝑙 − 𝐶𝑒1)

𝑀)

(4.2)

Where q2 is the solute equilibrium solid phase concentration after the second loading step

in mg solute/g NYEX®; q1 is the solute equilibrium solid phase concentration after the first

loading step in mg solute/g NYEX®; V is the liquid volume in L; Ccontrol is the solute

concentration in the control bottle in mg solute/L; Ce1 is the solute equilibrium liquid phase

concentration after the second loading step in mg solute/L; and Ce1 is the solute equilibrium

liquid phase concentration after the initial loading step in mg solute/L. Narbaitz and Cen

(1997) discussed there are alternative definitions that may provide more equitable

assessments of the regeneration efficiency and generate somewhat lower regeneration

efficiency values than the above. In this study, the above equation was utilized as it is the

most frequently used equation in the literature and thus provides a fairer comparison with

the work of others.

For the electrochemical regeneration over multiple cycles, several steps of electrochemical

regeneration and re-adsorption were repeated, after the final cycle the final adsorbent mass

was determined.

84

4.3.4 Atrazine Analysis

The atrazine analysis was performed using a modification of the HPLC-based direct water

injection methods developed by Agilent Technologies and the USGS (Fu 2008;

Steinheimer et al. 1990). These concentrations were determined using an Agilent 1100

series HPLC with a reversed phase C18 HyperSil GOLD column. The mobile flow rate

was 1 mL/min of a 55:45 CH3OH:H2O mixture with an intermediate injection blanks of

methanol and water between each sample. The detection limit of atrazine for this method

was 2 µg/L. Since environmental levels of atrazine in water generally range from non-

detectable to 10 µg/L (Duris et al. 2004; Gilliom et al. 2006), initially, the atrazine

analytical method involved a pre-concentration step using solid phase extraction (SPE)

with methanol elution which has a detection limit of this method was 0.4 µg/L (Steinheimer

et al. 1990). However, upon further examination of the stability of the chromatogram with

pure water solutions, it was determined that direct injection from the sample worked well

and the pre-concentration step was not necessary. The sensitivity of the HPLC analysis was

improved by increasing the injection volume to 100 µL. During the analysis of pure

atrazine, the method was working well. With the introduction of humic acids in order to

model the competitive adsorption behaviour, there were disruptions in the chromatograms.

In order to rectify the problem, potential solutions were examined, additional methanol

blanks in between sample injections eliminated the problems. Calibration curves were

prepared for atrazine in the 5 to 2000 µg/L range and it was confirmed using several

atrazine standards that also contained humic acid. The atrazine standards were prepared

using a 100 mg/L atrazine in methanol stock solution and diluted using a mechanical

pipette.

85

4.3.5 DOC Analysis

DOC analysis was performed to quantify the concentrations of humic acid. DOC was

analyzed using a UV-persulfate oxidation-based analyzer (Phoenix 8000, Tekmar-

Dohrmann, Cincinnati, OH) with a detection limit of 0.1 mg/L using Standard Methods

number 5310C, (APHA/AWWA/WEF 2017). Calibration was performed several

potassium phthalate standards in the 1 to 8 mg/L range. Higher concentration samples were

diluted. Prior to analysis all DOC samples were filtered through a 0.45 µm Whatman®

membrane filter.

4.1 Results and Discussion

4.1.1 Adsorption Studies of Humic Acid

Initially, the batch-point isotherms were performed using a constant concentration of humic

acid solution whilst changing the adsorbent mass of the NYEX®. As shown by the square

symbols in Figure 4.2, the constant initial solute concentration isotherm shows an

unexpected pattern. As all the equilibrium liquid phase concentrations were above 0.8 - 1

mg DOC/L, approximately 1 mg DOC/L of the humic acid was not adsorbable, which is

rather high but a fairly common phenomena in the adsorption of NOM solutions. Initially,

as expected, the adsorption capacity or loadings (i.e., equilibrium solid phase

concentrations) increases as the equilibrium liquid phase concentrations (Ceq) increases.

But as Ceq increases beyond 4 mg DOC/L, the loadings level off and then decrease. This

was puzzling as this pattern is typical of competitive adsorption, however this was for the

adsorption of a single solute. Ning (2015) observed a similar pattern, i.e., drop in capacity

for higher Ceq values, while conducting constant initial solute concentration bottle-point

isotherms for the adsorption of acid violet 17 onto NYEX® 1000. Through several

86

experiments he established the drop occurred irrespective of the initial solute

concentration. Ning (2015) speculated about potential reasons for the drop in capacity but

he did not identify the cause of the phenomena. Adsorption on NYEX® occurs primarily

on the particle surface (Asghar et al. 2014) so the NYEX® adsorption capacity is expected

to be greater for the smaller NYEX® particles. It should be noted that the NYEX® particles

are not very uniform in size. The lower loadings achieved at the higher equilibrium liquid

phase concentration, which correspond to the lower adsorbent doses, may be lower because

inadvertently large NYEX® particles were used. However, this seems unlikely because one

would expect that particle sizes used would be random which would not yield such

consistent pattern of decreasing loadings. An alternative potential reason for the decreasing

loadings is competitive adsorption with some unknown compound leaching from the

NYEX®. Brown et al. (2004a) observed a NYEX® leaching when they were conducting

atrazine experiments and the leachate was causing issues with developing calibration

curves. DOC analysis of the distilled water blanks of the current experiment identified there

was some leaching that ranged from 0.1-0.3 mg DOC/L.

Certainly, for the bottle points with the lowest adsorbent dosage there would be a greater

leaching driving force (i.e., the larger volume of solution per unit mass), so there could be

greater interference with the adsorption of the humic acid. Thus, it seems compound(s)

leaching from the NYEX® are the most likely explanation for the observed results, and

further research should be considered in identifying in quantifying the leaching of organics

from NYEX®. Because of the unusual results, the use of an alternative bottle-point isotherm

procedure was studied, i.e., the constant adsorbent dose isotherms approach.

87

Figure 4.2: NYEX® 1000 adsorption of humic acid using different adsorbent doses and

the same concentration solution.

Figure 4.3 shows the humic acid – NYEX® 1000 isotherm generated using a constant

adsorbent dose varying the concentration of adsorbate (shown as triangles) and for

comparison it also includes the constant solute concentration isotherm (shown as squares).

The change in the experimental procedure has a significant impact on the isotherm shape.

The constant adsorbent dose isotherm shows a clear increasing equilibrium loading with

increasing equilibrium liquid phase concentration pattern. Accordingly, the constant

adsorbent dose bottle-point isotherm method was used for all of the remaining isotherm

experiments. The data shows some scatter which is likely due to the lack of uniformity in

the NYEX®, this level of scatter has been observed by other researchers studying NYEX®

0

0.2

0.4

0.6

0.8

1

1.2

0 2 4 6 8 10 12 14

Equ

ilib

riu

m s

olid

ph

ase

co

nce

ntr

atio

n

(mgT

OC

/gN

YEX

)

Equilibrium liquid phase concentration (mgTOC/L)

Constant Ci of

Humics

88

adsorption and regeneration (Asghar et al. 2013). Due to the scatter in the constant dose

isotherm data it is not immediately evident if the can best described by a Langmuir or

Freundlich model.

Figure 4.3: Alternative NYEX® 1000 adsorption of humic acid isotherms: a) using a

constant dose of NYEX® and b) using a constant Ci of humic acid

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

0 2 4 6 8 10 12 14

Equ

ilib

riu

m s

olid

ph

ase

co

nce

ntr

atio

n

(mgT

OC

/gN

YEX

)

Equilibrium liquid phase concentration (mgTOC/L)

Constant Dose of

NYEX

Constant Ci of

Humics

89

Table 4.1 shows the results of linear regression of these two models. Based on the R2

coefficient the Freundlich model fits the data considerably better than the Langmuir model,

the relatively low R2 values are attributed to the scatter of the data. The average absolute

errors also indicate that the Freundlich model describes this data better. It should be noted

that Asghar et al. (2013) used humic acid concentrations directly instead of DOC in the

current study, so conversion was required to compare the data. In addition, our regression

analysis of the data in their graphs yielded different regressed coefficient values than those

reported in their tables. Using the DOC concentrations of our standard solutions we

obtained a DOC/humic acid ratio of 0.67, which was used to convert the two data sets on

the same basis. The best fit Freundlich exponent (1/n) in the current study was 0.49 and its

95% confidence limits was 0.354 to 0.623, these values overlap with the 0.495 + 0.0794

value obtained in the reanalyzed Asghar et al. (2013) data for the same adsorbate-adsorbent

combination. So, they are statistically the same. The values of KFREUND (at 0.658 ± 0.136

mgDOC/g NYEX®*(L/mg)1/n) is somewhat larger the 0.54 mg DOC/g NYEX®*(L/mg)1/n

converted/regressed value from based on the Asghar et al. (2013) data. Although the 95%

confidence limits of KFREUND overlap, the loadings in the current study are higher than

those reported by Asghar et al. (2013). Given that the humic acid sorption capacity of the

NYEX® 1000 used in this study are somewhat higher than those tested by Asghar et al.

(2013), the two batches of NYEX® 1000 are possibly somewhat different.

90

Table 4.1: Humic acid isotherm model coefficients and quality of fit.

Best fit &

95%

confidence

limits

Absolute

Percent

Error

R2

Freundlich

Model

KFREUND

[(μg DOC / g AC)/(μg DOC/L)1/n ]

0.658 ± 0.136

10.5 0.769

1/n 0.490 ± 0.133

Langmuir

Model

Qm 1.80 ± 0.698

13.7 0.4097

KLANG 0.708 ± 0.550

4.1.2 Adsorption Studies of Atrazine

The atrazine isotherm developed in Figure 4.4 shows a clear favourable adsorption trend

and the data shows less variability than that for the humic acid data. As expected, the

atrazine adsorption capacities on NYEX® 1000 is significantly smaller than those for

activated carbon. For example, at an equilibrium liquid phase concentration of 40 μg/L the

adsorption capacity for NYEX® 1000 is 3 µg atrazine/g, while from atrazine-PAC

isotherms in the literature the adsorption ranged from 3200 to 16200 µg atrazine/g (Ding

2010; Knappe et al. 1998; Li et al. 2003). The difference is even larger when comparing

this material with atrazine-GAC isotherms in the literature that can range from 56900 to

113600 µg atrazine/g (Schreiber et al. 2007; Sontheimer et al. 1988; Speth and Miltner

1990). Since the surface area of PAC and GAC is derived almost entirely from the internal

pore structure, it is accepted that the adsorption capacity of a GAC and the same GAC

ground to a powder is essentially the same. Accordingly, the 35.5-fold range of ATZ

loadings calculated from the isotherms or from graphs reported in the literature, is puzzling

91

and it seems to indicate some type of errors by some researchers. The figure also includes

the best fit Freundlich and Langmuir simulations as well as the data’s error bars estimated

using an estimated error analysis. The Freundlich and Langmuir regression analysis results

are presented in Table 4.2.

Table 4.2: Atrazine isotherm model coefficients and quality of fit.

Best fit &

95% confidence

limits

Average

Absolute

Error

R2

Freundlich

Model

KFREUND

[(μg ATZ / g AC)/(μg ATZ/L)1/n ]

0.147 ± 0.044

14.7 0.972

1/n 0.827 ± 0.0526

Langmuir

Model

Qm 49.2 ± 6.58

7.57 0.909

KLANG

1.74*10-3 ±

9.8*10-5

92

Figure 4.4: NYEX® 1000 single-solute atrazine adsorption isotherm

While the R2 values for the linear regression results presented in Table 4.2 seem to indicate

that the atrazine isotherm data is better described by the Freundlich model, however visual

observation of the simulations (Figure 4.4) and the average absolute error indicated

otherwise. The reason for this apparent contradiction seems to arise from log

transformations of the Freundlich model that is necessary to conduct a linear regression of

the data. Unfortunately, the atrazine adsorption data available in the literature is for

NYEX® 100 not for the NYEX® 1000 used in this study. The NYEX® 1000-atrazine

isotherm’s Freundlich constants (KFREUND = 0.147 (μg atrazine/g NYEX®)/((mg

0

5

10

15

20

25

30

35

40

0 200 400 600 800 1000

q (

µg

ATZ

/gN

YEX

)

Ce (µg ATZ/L)

Data

Freundlich Model

Langmuir Model

93

atrazine/L)0.827) and 1/n = 0.827) were significantly different from those found by Brown

et al. (2004) (KFREUND = 0.279 (μg atrazine/g NYEX®)/((mg atrazine/L)0.55) and 1/n =

0.550) for their atrazine-NYEX® 100 isotherm. The NYEX® 100 tested by Brown et al.

(2004) had a significantly higher atrazine sorption capacity, for example at an equilibrium

atrazine liquid phase concentration of 700 μg/L they found that NYEX® 100 has an atrazine

capacity of approximately 229 μg/g NYEX® 100 while in the current study the capacity is

approximately 27 μg/g NYEX® 1000.

Figure 4.5 shows that the presence of humic acid reduces the atrazine adsorption capacity,

for example at an equilibrium atrazine liquid phase concentration of 700 μg atrazine/L the

atrazine capacity decreases by approximately 41%. On average the competition with the

humic acids decreased atrazine sorption capacity by 37%. This could be due to the strong

bonds of the humics blocking the atrazine from being adsorbed on the adsorption site of

the NYEX®.

94

Figure 4.5: NYEX® adsorption isotherm for atrazine in the presence of humic acid

Unfortunately, the DOC concentrations for the above competitive adsorption experiments

were not performed. However, the impact of the competitive adsorption on the humic acid

adsorption can be assessed using the bi-solute loading results which are part of regeneration

experiments presented in the later parts of this manuscript. For the loading tests with the

closest initial humic acid concentration (i.e., 9.37 mg DOC/L) the equilibrium humic acids

loading was only 0.046 mg DOC/g NYEX® compared with 2.83 mg DOC/g NYEX®

predicted by the humic acid Freundlich isotherm. This represents a 98% decrease in the

humic acid capacity, which is a surprising finding for two reasons. First, in target-

0

5

10

15

20

25

30

35

0 200 400 600 800 1000

q (

µgA

TZ/g

NY

EX)

Ce (µgATZ/L)

Atrazine and Humics

Adsorption Isotherm

Atrazine (Langmuir)

Humic Acid Concentration

= 8 mg DOC/L

95

organics/NOM AC competitive adsorption tests the literature concentrates on the target

compound sorption only contains a limited number of reports on the adsorption of the

NOM in their experiments. Narbaitz and Benedek (2004) showed that competitive

adsorption with 1,1,2 trichloroethane did not significantly decrease the adsorption of a river

water NOM. Second, in the current NYEX® 1000 study the adsorption of atrazine

decreased the humic acid adsorption capacity by 98% while the humic acid only decreased

the atrazine adsorption capacity by an average 37%. So, the atrazine is the dominant solute.

This is contrary to that observed in target-organics/NOM competitive adsorption tests with

AC. Two common factors cited for the dominance of NOM in competitive adsorption on

AC are higher adsorbability of the NOM and much larger concentrations in the solutions

tested. Based on an Aldrich humic acids molecular weight of 4868 g/mol (Roger et al.

2010), the molar humic acids concentration in the adsorbate solutions was estimated to be

1.64 x 10-3 mmol/L. When the single solute isotherm data for humic acids and atrazine

isotherms developed above were compared on a molar basis, the humic acids had

significantly higher adsorption capacities, thus this was not factor in atrazine’s dominance

in the competitive adsorption tests. The initial atrazine concentrations of the various points

in Figure 4.5 were different in all cases. If the initial atrazine concentration and the initial

humic acids concentration are compared on a DOC basis, the humics are present in much

higher concentrations. This is the reasoning provided by other researchers for low impact

on the NOM adsorption capacities of competition with trace level contaminants (Ding

2010; Li et al. 2003) However, on a molar basis the initial atrazine concentration was

significantly higher than the humic acid initial concentration, so this possibly explains that

atrazine was the dominant solute in the above tests. It should be noted that most AC target-

96

organics/NOM competitive adsorption tests in the literature use the NOM present natural

waters which have average molecular weights around or below 1000 g/mol (Ding 2010; Li

et al. 2003; Smith 1991), and thus their initial NOM concentrations on a molar basis is

significantly higher than that of the target trace contaminant. In addition, given that the

adsorption sites on NYEX® 1000 are on the surface the dominance of atrazine may be

related to faster diffusion as it has a much higher liquid phase molecular diffusivity (6.5 x

10-10 m2/s) than the humic substances (1.88 x 10-10 m2/s) by a factor of 3.5 (Park et al. 2003;

Sontheimer et al. 1988). Given the unusual results obtained for the competitive adsorption

on NYEX®, these results should be verified and repeated for other target compounds and/or

atrazine and other humic substances. The next section will discuss the electrochemical

regeneration of the NYEX® 1000 testing various parameters.

4.1.3 Electrochemical Regeneration

4.1.3.1 Impact of Regeneration Time

The first set of regeneration tests investigated the impact of regeneration time at a constant

current density for NYEX® that was loaded with an atrazine-humic acid mixture. For these

regeneration tests a current density of 20 mA/cm2 was chosen based on Brown et. al

(2004a) who found it to yield the highest regeneration efficiency for atrazine-loaded

NYEX® 100. Figure 4.6 shows that the regeneration efficiency increased linearly for

regeneration times of 5 to 30 minutes and decreased somewhat for the one longer

regeneration time evaluated. Full regeneration was achieved with regeneration time of 20

minutes, and 10 additional minutes increased the efficiency to approximately 160%. This

high value was presumably achieved because the electrochemical regeneration modified

the surface of the NYEX® 1000. No comparable atrazine-NYEX® 1000 data is available in

97

the literature. For the regeneration of atrazine-loaded NYEX® 100 Brown et al. (2004a)

found that for a regeneration time of 10 minutes and a current density of 20 mA/cm2 (the

same as the current study) achieved over 100% RE. It is hypothesized, that the higher

regeneration time (30 min. vs. 10 min.) was necessary to regenerate the NYEX® 1000 due

to the presence of humic acid. There is the possibility that the enhanced adsorption occurred

during the second loading cycle because during the first regeneration some compounds

leached from the NYEX®.

Figure 4.6: Atrazine regeneration efficiency as a function of regeneration time

0

20

40

60

80

100

120

140

160

180

200

0 10 20 30 40 50 60

% R

E

TIME (MIN.)

Current Density = 20 mA/cm2

Atrazine concentration = 200 µg/L

Humic acid concentration = 7 mgDOC/L

98

4.1.3.2 Impact of current density on regeneration efficiency

Given that competitive adsorption with humic acid reduce the atrazine adsorption capacity,

it was speculated that the NYEX® anodic regeneration efficiency will be lower for the

NYEX® loaded with atrazine-humic mixture than that loaded with the single-solute

atrazine solution. The impact of current density on atrazine regeneration efficiency was

evaluated with NYEX® loaded with just atrazine and also with NYEX® loaded with

atrazine and humic acid (Figure 4.7).

For the atrazine loaded NYEX® (the squares in Figure 4.7) the regeneration efficiency

increased with increasing current density up to 20 mA/cm2, reaching a RE of approximately

140%. The regeneration efficiency remained constant at higher current densities. These

results are consistent with the findings of Brown et al. (2004) for the regeneration of

NYEX® 100 loaded with atrazine. They found that the optimal current density was 20

mA/cm2 and for higher values the RE decreased. They attributed this behaviour to side

reactions that would consume more electricity.

Tests were conducted to establish the impact of current density for the regeneration

efficiency of NYEX® loaded with an atrazine-humics solution showed a somewhat

different pattern (the triangles in Figure 4.7), the regeneration efficiency continued to

increase approximately linearly up to a current density of 30 mA/cm2. Accordingly, the

above mentioned side reactions do not seem to affect the NYEX® when humic acids are

present (Brown 1995). It is possible that the strong polarity of the humic acids is disrupting

the atrazine side reactions that occur at the increased current densities. Brown (1995)

hypothesized that this was the case in his atrazine experiments. This would mean that the

humic acids are actually benefitting the regeneration rather than being a hindrance.

99

Surprisingly, upon regeneration there was a slightly higher equilibrium atrazine loading

achieved by the NYEX® loaded with atrazine plus humic acid than by the NYEX® loaded

with just atrazine (15 µg atrazine/g NYEX® 1000 versus 13.1 µg atrazine/g NYEX® 1000).

This suggest the electrochemical regeneration of NYEX® minimally loaded with humic

acid helps improve the adsorbent to the point that compensates for the competitive

adsorption impact of the humics in subsequent loading cycles. This will be discussed in the

next section.

Figure 4.7: Atrazine RE efficiency versus current density for NYEX® 1000 loaded with

atrazine and humic acid and NYEX® 1000 leaded with atrazine

0

50

100

150

200

250

0 10 20 30 40 50

% R

E

CURRENT DENSITY (mA/cm2)

Atrazine

Atrazine & Humics

Atrazine Concentration: 200 µg/L

Humic Acid Concentration: 7

mgDOC/L

Regen time: 30 min.

100

This pattern in Figure 4.7 is consistent with results found by Asghar (2011) for the

adsorption of acid violet 17 onto NYEX® 1000 and its regeneration. For current densities

above 15 mA/cm2, the %RE flattened and did not increase above 85% (Asghar 2011).

4.1.3.3 Multiple Regeneration Cycles

Multiple cycle tests were performed in order to determine the consistency of the

electrochemical regenerations and to evaluate the long-term performance of the process.

Regeneration tests were conducted at the previously identified optimal conditions, a current

density of 20 mA/cm2 and a regeneration time of 30 minutes. These tests were performed

both with NYEX® loaded with atrazine and with NYEX® loaded with atrazine plus humic

acid. The initial solute concentrations were the same as in the previous tests. The

regeneration of NYEX® loaded with just atrazine, as shown as squares in Figure 4.8 was

120% in the first regeneration cycle and increased to 140% in the second and subsequent

cycles. The fact that the first regeneration cycle resulted in a regeneration over 100%

suggests the electrochemical process altered the surface of NYEX® surface. This possibly

resulted in the bonding of hydrogen ions to the NYEX® and improved it as an adsorbent of

negatively charged species. Atrazine behaves as a weak base in aqueous solutions and has

a pKa of 1.60 (NCBI 2004), thus at a neutral pH atrazine should be in its ionized form and

likely more attracted to the H+ ions on the NYEX® (Weber 1970, 1993; Weber et al. 1969).

Since humic acids are known to contain both positive and negative charges the positive

charges, this may result in the repulsion of the humic acids, thus lowering its adsorption

capacity on electrochemically regenerated NYEX®. In addition, the higher loading may

possibly be achieved by the availability of sorption sites that become available due to the

leaching of some organics originally on the NYEX®. Brown et al. (2004a) experienced

101

issues with the leaching of unknown compounds from the NYEX® that were making their

atrazine analysis difficult and the control samples in the current study also detected some

organic compound leaching for the NYEX® 1000. The consistent performance over the

next cycles showed consistent long-term performance. Brown et al. (2004a) also showed

that NYEX® 100 research which shows consistent atrazine regeneration efficiencies above

100%.

The regeneration of NYEX® loaded with atrazine plus humic acid, identified as the

triangles in Figure 4.8, showed a similar RE pattern but with somewhat higher RE values.

In this case, the long-term RE was approximately 170%. To confirm these higher values

were not an artifice of the RE formula, the data of these experiments was replotted in terms

of the equilibrium atrazine loading achieved in the multi regeneration cycle experiments

(Figure 4.9). This graph shows for the second and subsequent cycles achieved the same

atrazine loadings for both the atrazine-only solutions and the atrazine-humic acid solutions.

This implies that in these loading cycles the humic acids were no longer competing with

atrazine for the NYEX® sorption sites. And indeed, the adsorption of humic acids in the

second and subsequent cycles was negligible. It is hypothesized that the electrochemical

regeneration leaves a residual charge on the NYEX® and it minimizes the adsorption of

humic acids through charge repulsion. Future studies should attempt to confirm this

hypothesis. Regardless of the mechanism, it is evident that the RE graph (Figure 4.8) gives

a false impression in that the performance for the atrazine-humic acid mixture is superior

to that for the atrazine alone solutions. On the other hand, Figure 4.9 demonstrates that the

presence of humic-acids does not have a long-term detriment to the removal of atrazine by

NYEX® 1000 adsorption/oxidation systems. Presumably, the presence of the humic acid

102

during the first electrochemical regeneration cycle led to a slight change of the NYEX®

during this regeneration, and this change resulted in the higher atrazine sorption capacity

in later loading cycles. It is noteworthy that the electrochemical regeneration compensated

for the decrease in atrazine sorption capacity due to the competitive adsorption with humic

acid.

Figure 4.8: Impact of multiple regeneration cycles on the regeneration efficiency of

NYEX® loaded with atrazine and NYEX® loaded with atrazine and humic acid

0

20

40

60

80

100

120

140

160

180

200

0 1 2 3 4 5

% R

E

Regeneration Cycle

Atrazine

Atrazine & Humics

Atrazine Concentration: 200 µg/L

Humic Acid Concentration: 7 mgDOC/L

Regeneration time: 30 min.

Current Density = 20 mA/cm2

103

Figure 4.9: Multiple cycles of NYEX® regeneration using qee with atrazine and

atrazine & humic acid

4.4 Conclusions

The humic acid bottle-point isotherm conducted using the same solute and different

NYEX® 1000 doses yielded unexpected lower loadings for the smallest NYEX® dosages,

this is consistent with the findings of Ning (2015). The lower loadings could be caused by

competitive adsorption with organic compounds leached from the NYEX® 1000, this

requires further investigation (Ning 2015). Unfortunately, due to the electrostatic adhesive

characteristics of NYEX® it is very difficult to pre-wash NYEX® without the loss of the

smallest particles, which are those providing the higher surface area per unit mass. The

0

2

4

6

8

10

12

14

16

18

20

0 1 2 3 4 5

q (

µg/

g)

Regeneration Cycle

Atrazine

Atrazine & Humics

Atrazine Concentration: 200 µg/L

Humic Acid Concentration: 7 mgDOC/L

Regeneration time: 30 min.

Current Density = 20 mA/cm2

104

bottle-point method with a constant adsorbent dose yielded reasonable results and it is

recommended for experiments with NYEX®. The humic acid isotherm yielded similar

results to the data reported by Asghar et al. (2013). The humic acid single solute isotherm

was not as uniform as the atrazine isotherm showing great variability in the adsorption

behaviour of the humic acids. The atrazine-humic acid-NYEX® 1000 isotherms

demonstrated that, like with activated carbon, there is significant competitive adsorption

on NYEX® between the target trace toxic compound (atrazine) and humic acid. Since

NYEX® 1000 has virtually no internal pore structure, the competitive adsorption is not

related to pore blocking by the humic acids, which was one of the hypothesized

mechanisms for competitive adsorption onto activated carbon (Summers et al. 2012).

Surprisingly, the competition decreased the humic acids sorption capacity significantly

more (e.g. 98% decrease) than the atrazine capacity (average 37% decrease). It is

speculated that this was caused by: a) the initial atrazine concentrations (on a molar basis)

were significantly higher than those of the humic acids; and b) the diffusion of atrazine is

significantly faster.

The electrochemical regeneration of atrazine loaded NYEX® and NYEX® loaded with an

atrazine –humic acid solution was very effective. For several combinations of regeneration

time and current densities it achieved regeneration efficiencies of over 100%, presumably

the electrochemical processing altered the NYEX® improving its adsorption capacity.

However, the multi-cycle loading/electrochemical regeneration tests showed that NYEX®

loaded with atrazine-only solutions and the atrazine-humic acid solutions yielded the same

long-term atrazine loadings. Thus, the humic acid competition did not hinder NYEX®

105

1000's long-term removal of atrazine. These findings should be confirmed with other trace

organic contaminants and other sources of NOM.

106

References

APHA/AWWA/WEF. (2017). 5310 TOTAL ORGANIC CARBON (TOC) (2017). In

Standard Methods for the Examination of Water and Wastewater. Standard Methods

For the Examination of Water and Wastewater.

https://doi.org/10.2105/SMWW.2882.104

Asghar, H. M. A. (2011). Development of Graphitic Adsorbents for Water Treatment using

Adsorption and Electrochemical Regeneration. University of Manchester.

Asghar, H. M. A., Hussain, S. N., Roberts, E. P. L., & Brown, N. W. (2013). Removal of

humic acid from water using adsorption coupled with electrochemical regeneration.

Korean Journal of Chemical Engineering, 30(7), 1415–1422.

https://doi.org/10.1007/s11814-013-0066-z

Asghar, H. M. A., Hussain, S. N., Sattar, H., Roberts, E. P. L., & Brown, N. W. (2014).

Improved phenol adsorption from aqueous solution using electrically conducting

adsorbents. Korean Journal of Chemical Engineering, 31(5), 834–840.

https://doi.org/10.1007/s11814-013-0291-5

AWWA, & ASCE. (2013). ACTIVATED CARBON ADSORPTION. In Water Treatment

Plant Design (Fifth). Retrieved from

https://www.accessengineeringlibrary.com:443/browse/water-treatment-plant-

design-fifth-edition/c9780071745727ch16

Brown, N. W. (1995). Development of a cleaner process for the manufacture of exfoliating

graphite (p. 91; P. N. Sharratt, Ed.). p. 91. Manchester: UMIST.

107

Brown, N. W., & Roberts, E. P. L. (2007). Electrochemical pre-treatment of effluents

containing chlorinated compounds using an adsorbent. Journal of Applied

Electrochemistry, 37(11), 1329–1335. https://doi.org/10.1007/s10800-007-9376-3

Brown, N. W., Roberts, E. P. L., Chasiotis, A., Cherdron, T., & Sanghrajka, N. (2004).

Atrazine removal using adsorption and electrochemical regeneration. Water

Research, 38(13), 3067–3074. https://doi.org/10.1016/j.watres.2004.04.043

Brown, N. W., Roberts, E. P. L., Garforth, A. A., & Dryfe, R. A. W. (2004).

Electrochemical regeneration of a carbon-based adsorbent loaded with crystal violet

dye. Electrochimica Acta, 49(20), 3269–3281.

https://doi.org/10.1016/j.electacta.2004.02.040

Cannon, F. (1993). The effects of metals on thermal regeneration of granular activated

carbon (V. L. Snoeyink, Ed.). ProQuest Dissertations Publishing.

Chan, P.-Y., Lim, P.-E., Ng, S.-L., & Seng, C.-E. (2018). Bioregeneration of granular

activated carbon loaded with phenolic compounds: effects of biological and physico-

chemical factors. International Journal of Environmental Science and Technology,

15(8), 1699–1712. https://doi.org/10.1007/s13762-017-1527-4

Chin, Y., & Gschwend, P. M. (1991). The abundance, distribution, and configuration of

porewater organic colloids in recent sediments. Geochimica et Cosmochimica Acta,

55(5), 1309–1317. https://doi.org/10.1016/0016-7037(91)90309-S

Chowdhury, Z., Scott Summers, R., Westerhoff, G. P., Leto, B. J., Nowack, K. O., &

Corwin, C. J. (2013). Activated Carbon: Solutions for Improving Water Quality.

Retrieved from https://store.awwa.org/store/productdetail.aspx?productid=24610

108

Crittenden, J. C., Trussell, R. R., Hand, D. W., Howe, K. J., & Tchobanoglous, G. (2012).

MWH’s Water Treatment. In MWH’s Water Treatment.

https://doi.org/10.1002/9781118131473

Ding, L. (2010). Mechanisms of competitive adsorption between trace organic

contaminants and natural organic matter on activated carbon (University of Illinois

at Urbana-Champaign). Retrieved from

https://www.ideals.illinois.edu/handle/2142/16067

Duris, J. W., Reeves, H. W., & Kiesler, J. L. (2004). Atrazine Concentrations in Stream

Water and Streambed Sediment Pore Water in the St. Joseph and Galien River Basins,

Michigan and Indiana, May 2001 – September 2003. Reston: U.S. Geological Survey.

Fu, R. (2008). Analysis of Atrazine in Drinking Water at the ppb Level Using New Agilent

Reversed Phase LC Columns. Retrieved from

https://www.agilent.com/cs/library/applications/5989-8328EN.pdf

Gilliom, R. J., Barbash, J. E., Crawford, C. G., Hamilton, P. A., Martin, J. D., Nakagaki,

N., … Wolock, D. M. (2006). Pesticides in the Nation’s Streams and Ground Water,

1992–2001. In Circular (Revised Fe). https://doi.org/10.3133/cir1291

Hamdaoui, O., & Naffrechoux, E. (2007). Modeling of adsorption isotherms of phenol and

chlorophenols onto granular activated carbon. Part I. Two-parameter models and

equations allowing determination of thermodynamic parameters. Journal of

Hazardous Materials, 147(1–2), 381–394.

https://doi.org/10.1016/j.jhazmat.2007.01.021

Health Canada. (1993). Guidelines for Canadian Drinking Water Quality: Guideline

109

Technical Document - Atrazine. In Environmental and Workplace Health. Retrieved

from https://www.canada.ca/en/health-canada/services/publications/healthy-

living/guidelines-canadian-drinking-water-quality-guideline-technical-document-

atrazine.html#a1

Health Canada. (2015). Special Review of Paraquat : Proposed Decision for Consultation.

In Health Canada Pest Management Regulatory Agency. Retrieved from

https://www.canada.ca/content/dam/hc-sc/migration/hc-sc/cps-

spc/alt_formats/pdf/pest/part/consultations/rev2015-11/rev2015-11-eng.pdf

Hussain, S. N., de las Heras, N., Asghar, H. M. A., Brown, N. W., & Roberts, E. P. L.

(2014). Disinfection of water by adsorption combined with electrochemical treatment.

Water Research, 54, 170–178. https://doi.org/10.1016/j.watres.2014.01.043

Karimi-Jashni, A., & Narbaitz, R. M. (2005a). Electrochemical reactivation of granular

activated carbon: effect of electrolyte mixing. Journal of Environmental Engineering,

131(3), 443–449. https://doi.org/10.1061/(asce)0733-9372(2005)131:3(443)

Karimi-Jashni, A., & Narbaitz, R. M. (2005b). Electrochemical reactivation of granular

activated carbon: pH dependence. Journal of Environmental Engineering and

Science, 4(3), 187–194. https://doi.org/10.1139/s04-055

Kilduff, J. E., Karanfil, T., & Weber, W. J. (1996). Competitive interactions among

components of humic acids in granular activated carbon adsorption systems: effects

of solution chemistry. Environmental Science & Technology, 30(4), 1344–1351.

https://doi.org/10.1021/es950546z

Knappe, D. R. U., Matsui, Y., Snoeyink, V. L., Roche, P., Prados, M. J., & Bourbigot, M.-

110

M. (1998). Predicting the Capacity of Powdered Activated Carbon for Trace Organic

Compounds in Natural Waters. Environmental Science & Technology, 32(11), 1694–

1698. https://doi.org/10.1021/es970833y

Li, Q., Snoeyink, V. L., Mariãas, B. J., & Campos, C. (2003). Elucidating competitive

adsorption mechanisms of atrazine and NOM using model compounds. Water

Research, 37(4), 773–784. https://doi.org/10.1016/S0043-1354(02)00390-1

Liu, D., Roberts, E. P. L., Martin, A. D., Holmes, S. M., Brown, N. W., Campen, A. K., &

de las Heras, N. (2016). Electrochemical regeneration of a graphite adsorbent loaded

with Acid Violet 17 in a spouted bed reactor. Chemical Engineering Journal, 304, 1–

9. https://doi.org/10.1016/j.cej.2016.06.070

Lu, P. J., Lin, H. C., Yu, W. Te, & Chern, J. M. (2011). Chemical regeneration of activated

carbon used for dye adsorption. Journal of the Taiwan Institute of Chemical

Engineers, 42(2), 305–311. https://doi.org/10.1016/j.jtice.2010.06.001

McCreary, J. J., & Snoeyink, V. L. (1980). Characterization and activated carbon

adsorption of several humic substances. Water Research, 14(2), 151–160.

https://doi.org/10.1016/0043-1354(80)90231-6

Mohammed, F. M. (2011). Modelling and Design of Water Treatment Processes Using

Adsorption and Electrochemical Regeneration (University of Manchester,

Manchester, UK). Retrieved from https://www.escholar.manchester.ac.uk/uk-ac-

man-scw:128947

Moore, B. C., Cannon, F. S., Metz, D. H., & Demarco, J. (2003). GAC pore structure in

Cincinnati during full-scale treatment/reactivation. Journal - American Water Works

111

Association, 95(2), 103–112. https://doi.org/10.1002/j.1551-8833.2003.tb10296.x

Narbaitz, R. M., & Cen, J. (1994). Electrochemical regeneration of granular activated

carbon. Water Research, 28(8), 1771–1778. https://doi.org/10.1016/0043-

1354(94)90250-X

Narbaitz, R. M., & Karimi-Jashni, A. (2009). Electrochemical regeneration of granular

activated carbons loaded with phenol and natural organic matter. Environmental

Technology, 30(1), 27–36. https://doi.org/10.1080/09593330802422803

Narbaitz, R. M., & Karimi-Jashni, A. (2012). Electrochemical reactivation of granular

activated carbon: Impact of reactor configuration. Chemical Engineering Journal,

197, 414–423. https://doi.org/10.1016/j.cej.2012.05.049

Narbaitz, R. M., & McEwen, J. (2012). Electrochemical regeneration of field spent GAC

from two water treatment plants. Water Research, 46(15), 4852–4860.

https://doi.org/10.1016/j.watres.2012.05.046

NCBI. (2004). Atrazine. Retrieved May 4, 2017, from Pubchem Database website:

https://pubchem.ncbi.nlm.nih.gov/compound/2256

Ning, C. (2015). Adsorption equilibrium studies of acid violet 17 onto NYEX 1000. Ottawa.

Park, J. H., Feng, Y., Ji, P., Voice, T. C., & Boyd, S. A. (2003). Assessment of

bioavailability of soil-sorbed atrazine. Applied and Environmental Microbiology,

69(6), 3288–3298. https://doi.org/10.1128/AEM.69.6.3288-3298.2003

Pelekani, C., & Snoeyink, V. L. (2000). Competitive adsorption between atrazine and

methylene blue on activated carbon: The importance of pore size distribution. Carbon,

112

38(10), 1423–1436. https://doi.org/10.1016/S0008-6223(99)00261-4

Randtke, S. J., & Snoeyink, V. L. (1983). Evaluating GAC adsorptive capacity. American

Water Works Association, 75(8), 406–413. https://doi.org/10.1002/j.1551-

8833.1983.tb05177.x

Reid, P. M., Wilkinson, A. E., Tipping, E., & Jones, M. N. (1990). Determination of

molecular weights of humic substances by analytical (UV scanning)

ultracentrifugation. Geochimica et Cosmochimica Acta, 54(1), 131–138.

https://doi.org/10.1016/0016-7037(90)90201-U

Roger, G. M., Durand-Vidal, S., Bernard, O., Mériguet, G., Altmann, S., & Turq, P. (2010).

Characterization of humic substances and polyacrylic acid: A high precision

conductimetry study. Colloids and Surfaces A: Physicochemical and Engineering

Aspects, 356(1–3), 51–57. https://doi.org/10.1016/j.colsurfa.2009.12.029

San Miguel, G., Lambert, S. ., & Graham, N. J. . (2001). The regeneration of field-spent

granular-activated carbons. Water Research, 35(11), 2740–2748.

https://doi.org/10.1016/S0043-1354(00)00549-2

Schreiber, B., Schmalz, V., Brinkmann, T., & Worch, E. (2007). The effect of water

temperature on the adsorption equilibrium of dissolved organic matter and atrazine on

granular activated carbon. Environmental Science and Technology, 41(18), 6448–

6453. https://doi.org/10.1021/es070704+

Singh, S., & Yenkie, M. K. N. (2004). Competitive adsorption of some hazardous organic

pollutants from their binary and ternary solutions onto granular activated carbon

columns. Water, Air, and Soil Pollution, 156(1–4), 275–286.

113

https://doi.org/10.1023/B:WATE.0000036821.39350.5f

Smith, E. H. (1991). Evaluation of multicomponent adsorption equilibria for organic

mixtures onto activated carbon. Water Research, 25(2), 125–134.

https://doi.org/10.1016/0043-1354(91)90021-H

Sontheimer, H., Scott Summers, R., & Crittenden, J. C. (1988). Activated carbon for water

treatment (2nd ed.). Karlsruhe [Germany] : Denver, Colo.: DVGW-Forschungsstelle,

Engler-Bunte-Institut, Universitat Karlsruhe ; distributed in the USA by AWWA

Research Foundation.

Speth, T. F., & Miltner, R. J. (1990). Technical Note: Adsorption Capacity of GAC for

Synthetic Organics. Journal - American Water Works Association, 82(2), 72–75.

https://doi.org/10.1002/j.1551-8833.1990.tb06922.x

Steinheimer, T. R., Survey, U. S. G., & Ondrus, M. G. (1990). Liquid chromatographic

determination of atrazine and its degradation products in water.

https://doi.org/10.3133/wri894193

Summers, R. S., R. U. Knappe, D., & Snoeyink, V. L. (2012). Adsorption of Organic

Compounds by Activated Carbon. In J. K. Edzwald & American Water Works

Association (Eds.), Water quality & treatment: a handbook on drinking water (6th

ed.). New York: New York : McGraw-Hill, c2012.

Weber, J. B., Weed, S. B., & Ward, T. M. (1969). Adsorption of s -triazines by soil organic

matter. Weed Science, 17(4), 417–421. https://doi.org/10.1017/S0043174500054424

Weber, Jerome B. (1970). Mechanisms of adsorption of s-triazines by clay colloids and

114

factors affecting plant availability. In Single Pesticide Volume: The Triazine

Herbicides (pp. 93–130). https://doi.org/10.1007/978-1-4615-8464-3_5

Weber, Jerome B. (1993). Ionization and sorption of fomesafen and atrazine by soils and

soil constituents. Pesticide Science, 39(1), 31–38.

https://doi.org/10.1002/ps.2780390105

Yagmur, E., Turkoglu, S., Banford, A., & Aktas, Z. (2017). The relative performance of

microwave regenerated activated carbons on the removal of phenolic pollutants.

Journal of Cleaner Production, 149, 1109–1117.

https://doi.org/10.1016/j.jclepro.2017.02.188

Yeganeh, M. M., Kaghazchi, T., & Soleimani, M. (2006). Effect of raw materials on

properties of activated carbons. Chemical Engineering & Technology, 29(10), 1247–

1251. https://doi.org/10.1002/ceat.200500298

115

CHAPTER 5: Conclusion and Recommendations

5.1 Conclusions

The main objective of this thesis was to study the competitive adsorption of atrazine and

humic acid onto NYEX® 1000 and NYEX®’s electrochemical regeneration behaviour. The

conclusions of this thesis are as follows.

1. Based on the results of the humic acid – NYEX® isotherms, the constant solution-

varying adsorbent dose bottle-point procedure is not recommended for NYEX®

isotherms. Leaching of compounds from the NYEX® is plausible explanation for

the decrease in capacity at the higher equilibrium liquid phase concentration/lower

adsorbent dose data points. The constant dose method was chosen since the results

were not as affected by the leaching effect of the NYEX® 1000.

2. The coefficients of the Freundlich isotherm regressed from the NYEX® 1000-

humic acid isotherm data yielded were statistically the same as those derived from

those obtained from the data published by Asghar et al. (2013). The single-solute

atrazine onto NYEX® 1000 yielded significantly lower atrazine capacities than

those reported by Brown et al. (2004a) when adsorbing atrazine onto NYEX® 100,

however, NYEX® 1000 and NYEX® 100 are different materials.

3. The presence of humic acid depressed the NYEX®-atrazine adsorption capacity by

an average 37%, similar to what one would expect for the adsorption onto AC.

Since NYEX® lacks the large internal pore structure of GAC, pore blockage does

not seem to be a significant factor in the competitive adsorption decrease in

atrazine sorption capacity. Surprisingly, the adsorption capacity of humic acid

116

decreased by 98% when competing with atrazine. It is speculated that this was due

at least in part to the higher atrazine molar concentrations in the solution and higher

molecular diffusivity of atrazine. Therefore, atrazine was the dominant adsorbate.

4. Electrochemical regeneration of atrazine-loaded NYEX® 1000 and atrazine-humic

acid loaded NYEX® 1000 was very successful in that under many conditions

atrazine regeneration efficiencies over 100% were achieved. The electrochemical

processing appears to modify the NYEX® 1000 leading to higher atrazine sorption

capacities. For the conditions tested, increasing electrochemical regeneration

current density increased the atrazine regeneration efficiency atrazine-humic acid

loaded NYEX® 1000. However, it was slightly different for atrazine-only loaded

NYEX® 1000 in that the regeneration efficiency decreased slightly for current

densities above 20 mA/cm2. Brown (1995) suggested that in the regeneration of

the atrazine-only loaded NYEX® at higher currents there were current-consuming

side reactions, it is speculated that humic acids disrupted these reactions.

5. Multiple loading/regeneration NYEX® test cycles showed that the presence of

humic acid appeared to improve the NYEX®’1000s atrazine sorption capacity so

that it compensates for decrease in atrazine sorption capacity caused by

competitive adsorption. It is speculated that a residual positive charge left on the

NYEX® 1000 turned its surface acidic, that could explain the attraction of a weak

base such as atrazine.

117

5.2 Recommendations

1. The variable dose experiments with NYEX® 1000 requires more thorough

examination, particularly trying to identify compounds leaching from the

adsorbent.

2. To reduce the electrostatic adhesion problems in the transfer of the NYEX®

particles, it is recommended that future NYEX® adsorption/electrochemical

regeneration be performed in the same reactor, such as that developed by Roberts

and colleagues. This will also help reduce the time required to conduct the

experiments. Liu et al. (2016) used one such reactor.

3. To Pre-treat the NYEX® through electrochemical regeneration to enhance the

surface of the NYEX® in order to study the impact of electrochemical processing

on the adsorption of humic acid.

4. The NYEX® competitive adsorption findings of this study should be confirmed

using other trace organic contaminants and other sources of NOM.

5. When increasing current density for regeneration sorbed atrazine onto NYEX®

1000, there is a decrease in regeneration efficiencies. This is not the case in the

sorption of atrazine & humics onto NYEX® 1000 where there was only an increase

in regeneration efficiency as current density increases. It was speculated this was

caused by side reactions involved that are causing the decrease in regeneration

efficiency and why these side reactions do not affect the multi-component

regeneration. Therefore, further research is needed in order to identify the specific

side reactions.

118

6. The multiple cycles of regeneration showed very little change in the regeneration

efficiencies using the same NYEX® 1000 batch of both the single solute and multi-

solute regenerations after 5 regeneration cycles. Therefore, it is worth investigating

if the NYEX® 1000 can sustain these results on a long-term basis. This would

particularly worthwhile if there is deterioration of the adsorbent after a large

number of regeneration cycles.

119

References

Ania, C. O., Parra, J. B., Menéndez, J. A., and Pis, J. J. (2005). “Effect of microwave and

conventional regeneration on the microporous and mesoporous network and on the

adsorptive capacity of activated carbons.” Microporous and Mesoporous Materials,

85(1–2), 7–15.

APHA/AWWA/WEF. (2017). “5310 TOTAL ORGANIC CARBON (TOC) (2017).”

Standard Methods For the Examination of Water and Wastewater, Standard Methods

for the Examination of Water and Wastewater, American Public Health Association.

Asghar, H. M. A. (2011). “Development of Graphitic Adsorbents for Water Treatment

using Adsorption and Electrochemical Regeneration.” University of Manchester.

Asghar, H. M. A., Hussain, S. N., Brown, N. W., and Roberts, E. P. L. (2019).

“Comparative adsorption–regeneration performance for newly developed

carbonaceous adsorbent.” Journal of Industrial and Engineering Chemistry, The

Korean Society of Industrial and Engineering Chemistry, 69, 90–98.

Asghar, H. M. A., Hussain, S. N., Roberts, E. P. L., and Brown, N. W. (2013). “Removal

of humic acid from water using adsorption coupled with electrochemical

regeneration.” Korean Journal of Chemical Engineering, 30(7), 1415–1422.

Asghar, H. M. A., Hussain, S. N., Sattar, H., Brown, N. W., and Roberts, E. P. L. (2015).

“Mercaptan’s removal from aqueous solution using modified graphite-based

adsorbent through batch-wise adsorption–regeneration.” Chemical Engineering

Communications, 2015, 202(9), 1155–1160.

120

Asghar, H. M. A., Hussain, S. N., Sattar, H., Roberts, E. P. L., and Brown, N. W. (2014).

“Improved phenol adsorption from aqueous solution using electrically conducting

adsorbents.” Korean Journal of Chemical Engineering, 31(5), 834–840.

Australian Government Department of Environment and Energy. (2019). “Phenol.”

National Pollutant Inventory, <http://www.npi.gov.au/resource/phenol> (Sep. 20,

2003).

AWWA, and ASCE. (2013). “ACTIVATED CARBON ADSORPTION.” Water

Treatment Plant Design, McGraw Hill Professional, Access Engineering, New York.

Ayotamuno, M. J., Kogbara, R. B., Ogaji, S. O. T., and Probert, S. D. (2006). “Petroleum

contaminated ground-water: Remediation using activated carbon.” Applied Energy,

83(11), 1258–1264.

Bakke, J. E., Larson, J. D., and Price, C. E. (1972). “Metabolism of atrazine and 2-

hydroxyatrazine by the rat.” Journal of agricultural and food chemistry, 20(3), 602–

7.

Bard, A. J., and Faulkner, L. R. (2001). Electrochemical methods : fundamentals and

applications. Wiley, New York.

Betancur-Corredor, B., Soltan, J., and Peñuela, G. A. (2016). “Mineralization of ibuprofen

and humic acid through catalytic ozonation.” Ozone: Science & Engineering, Taylor

& Francis, 38(3), 203–210.

Brown, N. W. (1995). “Development of a cleaner process for the manufacture of

exfoliating graphite.” (P. N. Sharratt, ed.), UMIST, Manchester.

121

Brown, N. W., and Roberts, E. P. L. (2007). “Electrochemical pre-treatment of effluents

containing chlorinated compounds using an adsorbent.” Journal of Applied

Electrochemistry, 37(11), 1329–1335.

Brown, N. W., Roberts, E. P. L., Chasiotis, A., Cherdron, T., and Sanghrajka, N. (2004a).

“Atrazine removal using adsorption and electrochemical regeneration.” Water

Research, 38(13), 3067–3074.

Brown, N. W., Roberts, E. P. L., Garforth, A. A., and Dryfe, R. A. W. (2004b).

“Electrochemical regeneration of a carbon-based adsorbent loaded with crystal violet

dye.” Electrochimica Acta, 49(20), 3269–3281.

Cannon, F. (1993). “The effects of metals on thermal regeneration of granular activated

carbon.” (V. L. Snoeyink, ed.), ProQuest Dissertations Publishing.

Cannon, F. S., Snoeyink, V. L., Lee, R. G., and Dagois, G. (1997). “Effect of iron and

sulfur on thermal regeneration of GAC.” Journal / American Water Works

Association, 89(11), 111–122.

CDC. (2014). “The Safe Water System.” Centres for Disease Control and Prevention

(CDC), <https://www.cdc.gov/safewater/chlorination-byproducts.html> (Dec. 2,

2016).

Chan, P.-Y., Lim, P.-E., Ng, S.-L., and Seng, C.-E. (2018). “Bioregeneration of granular

activated carbon loaded with phenolic compounds: effects of biological and physico-

chemical factors.” International Journal of Environmental Science and Technology,

15(8), 1699–1712.

122

Chen, D. Z., Zhang, J. X., and Chen, J. M. (2010). “Adsorption of methyl tert-butyl ether

using granular activated carbon: Equilibrium and kinetic analysis.” International

Journal of Environmental Science and Technology, 7(2), 235–242.

Chin, Y., and Gschwend, P. M. (1991). “The abundance, distribution, and configuration of

porewater organic colloids in recent sediments.” Geochimica et Cosmochimica Acta,

55(5), 1309–1317.

Chowdhury, Z., Scott Summers, R., Westerhoff, G. P., Leto, B. J., Nowack, K. O., and

Corwin, C. J. (2013). Activated Carbon: Solutions for Improving Water Quality.

American Water Works Association.

Cohen, S. Z., Creeger, S. M., Carsel, R. F., and Enfield, C. G. (2009). Potential Pesticide

Contamination of Groundwater from Agricultural Uses.

Cooney, D. O. (1999). Adsorption design for wastewater treatment. Boca Raton, Fl. :

Lewis Publishers, c1999., Boca Raton, Fl.

Cooney, D. O., Nagerl, A., and Hines, A. L. (1983). “Solvent regeneration of activated

carbon.” Water Research, 17(4), 403–410.

Crittenden, J. C., Trussell, R. R., Hand, D. W., Howe, K. J., and Tchobanoglous, G. (2012).

MWH’s Water Treatment. MWH’s Water Treatment, John Wiley & Sons, Inc.,

Hoboken, NJ, USA.

Crosignani, P., Berrino, F., Donna, A., Betta, P.-G., Robutti, F., and Bellingeri, D. (1984).

“Ovarian mesothelial tumors and herbicides: A case-control study.” Carcinogenesis,

5(7), 941–942.

123

Ding, L. (2010). “Mechanisms of competitive adsorption between trace organic

contaminants and natural organic matter on activated carbon.” University of Illinois

at Urbana-Champaign.

Dobbs, R. A., and Cohen, J. M. (1980). Carbon adsorption isotherms for toxic organics.

Municipal Environmental Research Laboratory, Office of Research and

Development, U.S. Environmental Protection Agency ; Available to the public

through the National Technical Information Service, Cincinnati, Ohio; Springfield,

Va.

Donna, A., Crosignani, P., Robutti, F., Betta, P. G., Bocca, R., Mariani, N., Ferrario, F.,

Fissi, R., and Berrino, F. (1989). “Triazine herbicides and ovarian epithelial

neoplasms.” Scandinavian Journal of Work, Environment & Health, 15(1), 47–53.

Duris, J. W., Reeves, H. W., and Kiesler, J. L. (2004). Atrazine Concentrations in Stream

Water and Streambed Sediment Pore Water in the St. Joseph and Galien River Basins,

Michigan and Indiana, May 2001 – September 2003. U.S. Geological Survey, Reston.

Fay, M., Brattin, W. J., and Donohue, J. M. (1999). “Public Health Statement.” Toxicology

and Industrial Health, Agency for Toxic Substances and Disease Registry (ATSDR),

15(8), 652–654.

Foo, K. Y., and Hameed, B. H. (2012). “A cost effective method for regeneration of durian

shell and jackfruit peel activated carbons by microwave irradiation.” Chemical

Engineering Journal, Elsevier B.V., 193–194, 404–409.

Fu, R. (2008). Analysis of Atrazine in Drinking Water at the ppb Level Using New Agilent

Reversed Phase LC Columns. Shanghai.

124

García-Otón, M., Montilla, F., Lillo-Ródenas, M. A., Morallón, E., and Vázquez, J. L.

(2005). “Electrochemical regeneration of activated carbon saturated with toluene.”

Journal of Applied Electrochemistry, 35(3), 319–325.

Gilliom, R. J., Barbash, J. E., Crawford, C. G., Hamilton, P. A., Martin, J. D., Nakagaki,

N., Nowell, L. H., Scott, J. C., Stackelberg, P. E., Thelin, G. P., and Wolock, D. M.

(2006). Pesticides in the Nation’s Streams and Ground Water, 1992–2001. Circular,

Reston, VA.

Guedidi, H., Reinert, L., Lévêque, J.-M., Soneda, Y., Bellakhal, N., and Duclaux, L.

(2013). “The effects of the surface oxidation of activated carbon, the solution pH and

the temperature on adsorption of ibuprofen.” Carbon, 54, 432–443.

Hamdaoui, O., and Naffrechoux, E. (2007). “Modeling of adsorption isotherms of phenol

and chlorophenols onto granular activated carbon. Part I. Two-parameter models and

equations allowing determination of thermodynamic parameters.” Journal of

hazardous materials, 147(1–2), 381–94.

Health Canada. (1993). Guidelines for Canadian Drinking Water Quality: Guideline

Technical Document - Atrazine. Environmental and Workplace Health, Ottawa.

Health Canada. (2015). Special Review of Paraquat : Proposed Decision for Consultation.

Health Canada Pest Management Regulatory Agency, Ottawa.

Himmelstein, K. J., Fox, R. D., and Winter, T. H. (1973). “In-place regeneration of

activated carbon.” Chemical Engineering Progress, 69(11), 65–69.

Hussain, S. N., Asghar, H. M. A., Sattar, H., Brown, N. W., and Roberts, E. P. L. (2015).

125

“Chlorinated breakdown products formed during oxidation of adsorbed phenol by

electrochemical regeneration of a graphite intercalation compound.” Journal of

Industrial and Engineering Chemistry, The Korean Society of Industrial and

Engineering Chemistry, 30, 212–219.

Hussain, S. N., de las Heras, N., Asghar, H. M. A., Brown, N. W., and Roberts, E. P. L.

(2014). “Disinfection of water by adsorption combined with electrochemical

treatment.” Water Research, Elsevier Ltd, 54, 170–178.

IARC. (1991). “Atrazine.” IARC monographs on the evaluation of carcinogenic risks to

humans, 53, 441.

Karimi-Jashni, A., and Narbaitz, R. M. (2005a). “Electrochemical reactivation of granular

activated carbon: effect of electrolyte mixing.” Journal of Environmental

Engineering, 131(3), 443–449.

Karimi-Jashni, A., and Narbaitz, R. M. (2005b). “Electrochemical reactivation of granular

activated carbon: pH dependence.” Journal of Environmental Engineering and

Science, 4(3), 187–194.

Kilduff, J. E., Karanfil, T., and Weber, W. J. (1996). “Competitive interactions among

components of humic acids in granular activated carbon adsorption systems: effects

of solution chemistry.” Environmental Science & Technology, 30(4), 1344–1351.

Knappe, D. R. U., Matsui, Y., Snoeyink, V. L., Roche, P., Prados, M. J., and Bourbigot,

M.-M. (1998). “Predicting the Capacity of Powdered Activated Carbon for Trace

Organic Compounds in Natural Waters.” Environmental Science & Technology,

32(11), 1694–1698.

126

Kniewald, J., Osredečki, V., Gojmerac, T., Zechner, V., and Kniewald, Z. (1995). “Effect

of s-triazine compounds on testosterone metabolism in the rat prostate.” Journal of

Applied Toxicology, 15(3), 215–218.

Lee, J.-W., Yang, T.-H., Shim, W.-G., Kwon, T.-O., and Moon, I.-S. (2007). “Equilibria

and dynamics of liquid-phase trinitrotoluene adsorption on granular activated carbon:

Effect of temperature and pH.” Journal of Hazardous Materials, 141(1), 185–192.

Leng, C.-C., and Pinto, N. G. (1996). “An investigation of the mechanisms of chemical

regeneration of activated carbon.” Industrial & Engineering Chemistry Research,

35(1994), 2024–2031.

Li, F., Yuasa, A., Ebie, K., Azuma, Y., Hagishita, T., and Matsui, Y. (2002). “Factors

affecting the adsorption capacity of dissolved organic matter onto activated carbon:

modified isotherm analysis.” Water research, 36(18), 4592–604.

Li, Q., Snoeyink, V. L., Mariãas, B. J., and Campos, C. (2003). “Elucidating competitive

adsorption mechanisms of atrazine and NOM using model compounds.” Water

Research, 37(4), 773–784.

Liu, D., Roberts, E. P. L., Martin, A. D., Holmes, S. M., Brown, N. W., Campen, A. K.,

and de las Heras, N. (2016). “Electrochemical regeneration of a graphite adsorbent

loaded with Acid Violet 17 in a spouted bed reactor.” Chemical Engineering Journal,

Elsevier B.V., 304, 1–9.

Lu, P. J., Lin, H. C., Yu, W. Te, and Chern, J. M. (2011). “Chemical regeneration of

activated carbon used for dye adsorption.” Journal of the Taiwan Institute of Chemical

Engineers, Taiwan Institute of Chemical Engineers, 42(2), 305–311.

127

McCrae, B. (1991). “The characterization and identification of potentially leachable

pesticides and areas vulnerable to groundwater contamination by pesticides in

Canada.” Agriculture Canada, Ottawa, 89(01).

McCreary, J. J., and Snoeyink, V. L. (1980). “Characterization and activated carbon

adsorption of several humic substances.” Water Research, 14(2), 151–160.

McLaughlin, H. (2005). “Understanding activated carbon reactivation and low-

temperature regeneration technology.” International Sugar Journal, 107, 112–127.

Mohammed, F. M. (2011). “Modelling and Design of Water Treatment Processes Using

Adsorption and Electrochemical Regeneration.” University of Manchester,

Manchester, UK.

Moore, B. C., Cannon, F. S., Metz, D. H., and Demarco, J. (2003). “GAC pore structure in

Cincinnati during full-scale treatment/reactivation.” Journal - American Water Works

Association, 95(2), 103–112.

Narbaitz, R. M. (2019). “Water and wastewater treatment with activated carbon.” Dept. of

Civil Engineering, Ottawa.

Narbaitz, R. M., and Benedek, A. (1994). “Adsorption of 1,1,2‐Trichloroethane from River

Water.” Journal of Environmental Engineering, American Society of Civil Engineers,

120(6), 1400–1415.

Narbaitz, R. M., and Cen, J. (1994). “Electrochemical regeneration of granular activated

carbon.” Water Research, 28(8), 1771–1778.

Narbaitz, R. M., and Cen, J. (1997). “Alternative methods for determining the percent

128

regeneration of activated carbon.” Water Research, 31, 2532–2542.

Narbaitz, R. M., and Karimi-Jashni, A. (2009). “Electrochemical regeneration of granular

activated carbons loaded with phenol and natural organic matter.” Environmental

Technology, 30(1), 27–36.

Narbaitz, R. M., and Karimi-Jashni, A. (2012). “Electrochemical reactivation of granular

activated carbon: Impact of reactor configuration.” Chemical Engineering Journal,

Elsevier B.V., 197, 414–423.

Narbaitz, R. M., and McEwen, J. (2012). “Electrochemical regeneration of field spent GAC

from two water treatment plants.” Water Research, Elsevier Ltd, 46(15), 4852–4860.

NCBI. (2004). “Atrazine.” Pubchem Database, National Center for Biotechnology

Information (NCBI), <https://pubchem.ncbi.nlm.nih.gov/compound/2256> (May 4,

2017).

Ning, C. (2015). Adsorption equilibrium studies of acid violet 17 onto NYEX 1000. Ottawa.

de Oliveira Pimenta, A. C., and Kilduff, J. E. (2006). “Oxidative coupling and the

irreversible adsorption of phenol by graphite.” Journal of Colloid and Interface

Science, 293(2), 278–289.

Park, J. H., Feng, Y., Ji, P., Voice, T. C., and Boyd, S. A. (2003). “Assessment of

bioavailability of soil-sorbed atrazine.” Applied and Environmental Microbiology,

69(6), 3288–3298.

Pelekani, C., and Snoeyink, V. L. (2000). “Competitive adsorption between atrazine and

methylene blue on activated carbon: The importance of pore size distribution.”

129

Carbon, 38(10), 1423–1436.

Radian, A., and Mishael, Y. (2012). “Effect of humic acid on pyrene removal from water

by polycation-clay mineral composites and activated carbon.” Environmental Science

& Technology, 46(11), 6228–6235.

Randtke, S. J., and Snoeyink, V. L. (1983). “Evaluating GAC adsorptive capacity.”

American Water Works Association, 75(8), 406–413.

Reid, P. M., Wilkinson, A. E., Tipping, E., and Jones, M. N. (1990). “Determination of

molecular weights of humic substances by analytical (UV scanning)

ultracentrifugation.” Geochimica et Cosmochimica Acta, 54(1), 131–138.

Rodríguez-González, N., Beceiro-González, E., González-Castro, M. J., and Muniategui-

Lorenzo, S. (2013). “Application of a developed method for the extraction of triazines

in surface waters and storage prior to analysis to seawaters of galicia (Northwest

Spain).” The Scientific World Journal, 2013, 1–6.

Roger, G. M., Durand-Vidal, S., Bernard, O., Mériguet, G., Altmann, S., and Turq, P.

(2010). “Characterization of humic substances and polyacrylic acid: A high precision

conductimetry study.” Colloids and Surfaces A: Physicochemical and Engineering

Aspects, 356(1–3), 51–57.

San Miguel, G., Lambert, S. ., and Graham, N. J. . (2001). “The regeneration of field-spent

granular-activated carbons.” Water Research, 35(11), 2740–2748.

Schreiber, B., Schmalz, V., Brinkmann, T., and Worch, E. (2007). “The effect of water

temperature on the adsorption equilibrium of dissolved organic matter and atrazine on

130

granular activated carbon.” Environmental Science and Technology, 41(18), 6448–

6453.

Shi, H., Ding, J., Timmons, T., and Adams, C. (2012). “pH effects on the adsorption of

saxitoxin by powdered activated carbon.” Harmful Algae, Elsevier B.V., 19, 61–67.

Sillanpaa, M. E. T., and Shestakova, M. (2017). Electrochemical water treatment methods :

fundamentals, methods and full scale applications. (M. Shestakova, ed.), Butterworth-

Heinemann, an imprint of Elsevier, 2017, Oxford.

Singh, S., and Yenkie, M. K. N. (2004). “Competitive adsorption of some hazardous

organic pollutants from their binary and ternary solutions onto granular activated

carbon columns.” Water, Air, and Soil Pollution, 156(1–4), 275–286.

Smith, E. H. (1991). “Evaluation of multicomponent adsorption equilibria for organic

mixtures onto activated carbon.” Water Research, 25(2), 125–134.

Sontheimer, H., Scott Summers, R., and Crittenden, J. C. (1988). Activated carbon for

water treatment. DVGW-Forschungsstelle, Engler-Bunte-Institut, Universitat

Karlsruhe ; distributed in the USA by AWWA Research Foundation, Karlsruhe

[Germany] : Denver, Colo.

Speth, T. F., and Miltner, R. J. (1990). “Technical Note: Adsorption Capacity of GAC for

Synthetic Organics.” Journal - American Water Works Association, John Wiley &

Sons, Ltd, 82(2), 72–75.

Steinheimer, T. R., Survey, U. S. G., and Ondrus, M. G. (1990). Liquid chromatographic

determination of atrazine and its degradation products in water.

131

Summers, R. S., R. U. Knappe, D., and Snoeyink, V. L. (2012). “Adsorption of Organic

Compounds by Activated Carbon.” Water quality & treatment: a handbook on

drinking water, J. K. Edzwald and American Water Works Association, eds., New

York : McGraw-Hill, c2012., New York.

Sun, Y., Zhang, B., Zheng, T., and Wang, P. (2017). “Regeneration of activated carbon

saturated with chloramphenicol by microwave and ultraviolet irradiation.” Chemical

Engineering Journal, Elsevier B.V., 320, 264–270.

Timchalk, C., Dryzga, M. D., Langvardt, P. W., Kastl, P. E., and Osborne, D. W. (1990).

“Determination of the effect of tridiphane on the pharmacokinetics of [14C]-atrazine

following oral administration to male fischer 344 rats.” Toxicology, 61(1), 27–40.

USEPA. (1999). Enhanced Coagulation and Enhanced Precipitative Softening Guidance

Manual. EPA 815-R-99-012, United States Enviromental Protection Agency

(USEPA).

Van Vliet, B. M. (1991). “The regeneration of activated carbon.” Journal of the Southern

African Institute of Mining and Metallurgy, 91(5), 159–167.

Wang, L., and Balasubramanian, N. (2009). “Electrochemical regeneration of granular

activated carbon saturated with organic compounds.” Chemical Engineering Journal,

155(3), 763–768.

Ward, T. M., and Getzen, F. M. (1970). “Influence of pH on the adsorption of aromatic

acids on activated carbon.” Environmental Science & Technology, 4(1), 64–67.

Weber, J. B. (1970). “Mechanisms of adsorption of s-triazines by clay colloids and factors

132

affecting plant availability.” Single Pesticide Volume: The Triazine Herbicides,

Springer International Publishing, New York, NY, 93–130.

Weber, J. B. (1993). “Ionization and sorption of fomesafen and atrazine by soils and soil

constituents.” Pesticide Science, 39(1), 31–38.

Weber, J. B., Weed, S. B., and Ward, T. M. (1969). “Adsorption of s -triazines by soil

organic matter.” Weed Science, 17(4), 417–421.

Yagmur, E., Turkoglu, S., Banford, A., and Aktas, Z. (2017). “The relative performance of

microwave regenerated activated carbons on the removal of phenolic pollutants.”

Journal of Cleaner Production, Elsevier Ltd, 149, 1109–1117.

Yang, C. (1998). “Statistical mechanical study on the freundlich isotherm equation.”

Journal of Colloid and Interface Science, 208(2), 379–387.

Yeganeh, M. M., Kaghazchi, T., and Soleimani, M. (2006). “Effect of raw materials on

properties of activated carbons.” Chemical Engineering & Technology, 29(10), 1247–

1251.

Yuen, F. K., and Hameed, B. H. (2009). “Recent developments in the preparation and

regeneration of activated carbons by microwaves.” Advances in Colloid and Interface

Science, Elsevier B.V., 149(1–2), 19–27.

Zhang, H., Ye, L., and Zhong, H. (2002). “Regeneration of phenol-saturated activated

carbon in an electrochemical reactor.” Journal of Chemical Technology &

Biotechnology, 77(11), 1246–1250.

Zhang, X., Ding, Z., Yang, J., Cizmas, L., Lichtfouse, E., and Sharma, V. K. (2018).

133

“Efficient microwave degradation of humic acids in water using persulfate and

activated carbon.” Environmental Chemistry Letters, Springer International

Publishing, 16(3), 1069–1075.

134

Appendix A

Comments on the Atrazine Analytical Method Development

This Appendix is to highlight additional subject matter required during the method

development phase for the analysis of atrazine and atrazine & humic acid.

Firstly, the retention time could be increased by raising the percent amount of methanol in

the mobile phase to 90%. The likely result of this approach is that the distortion would still

occur and the atrazine can be recovered at a later retention time. The main issue with this

method is that the consumption of methanol would potentially be greater. For instance, the

amount of methanol will be increased by 35% to 90%. With an average operating time of

20 hours and a mobile flow rate of 1 mL/min, the overall consumption of methanol would

be 1,080 mL of methanol. Comparing this scenario using 55% methanol the consumption

would be 660 mL. That is an overall increase of 420 ml for every 20 hours of operation.

This method would constitute a greater increase in the consumption of reagent which would

be more efficient in the consumption of reagents during the HPLC analysis.